AD-AI35 103 THE BASIC ORGANIZING/OPTIMIZING ...20. Al.TS ACT (Con nue on revere. ci e If neceey "d...

L AD-AI35 103 THE BASIC ORGANIZING/OPTIMIZING TRAINING SCHEDULER(BOOTS): SYSTEM OVERVIEW(U) TRAINING ANALYSIS AND

IEVALUATI[ON GROUP (NAVY) ORLANDO FL F L KEELER ET AL.UCISIID SEP 83 TAEG-TR-150 FIG 5/9 N

- Ll 32

'"'I I . f.. ~ l1~2.0

iiiiiU 1.8

M~ N)~)PIRL~L Tl , ',UHARI

TECHNICAL REPORT 150

THE BASIC ORGANIZING/OPTIMIZING TRAINING

N SCHEDULER (BOOTS):SYSTEM OVERVIEW

SEPTEMBER 1953

FOCUS ON THE T

DTJCEECTE

LU D

A ' PPROVED FOR PUBLIC RELEASE,;Lu DISTRIBUTIONISULMTD

Is UNLIMITED=

Technical Report 150

THE BASIC ORGANIZING/OPTIMIZING TRAINING SCHEDULER (BOOTS):SYSTEM OVERVIEW

F. Laurence KeelerRichard L. Church

Training Analysis and Evaluation Group

September 1983

NTIS GR&I

DTIC TABUnannounced [

GOVERNMENT RIGHTS IN DATA STATEMENT Justificatio-

Reproduction of this publication in whole By

or in part is permitted for any purpose Distribution/of the United States Government. Availability Co1es

Dist Special

ALFRED F. SMODE, Ph.D., DireLtor W. L. MALOY, Ed.D.Training Analysis and Evaluation Group Principal Civilian Advo

on Education and Training

,r............ ............. ..... ........... ...... ......... ,... ,.liB.I.....In.........---

Unclassified

SECURITY CLASSIFICATION OF THIS PAGE (Wh n D.* Entered)

REPORT DOCUMENTATION PAGE READ INSTRUCTIONSREPORTDOCUMENTATIONPAGE_ BEFORE COMPLETING FORM

I. REPORT NUMBER 2. GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER

Technical Report 150 3)/ < 14. TITLE (snd Subtitle) S. TYPE OF REPORT & PERIOD COVERED

THE BASIC ORGANIZING/OPTIMIZING TRAININGSCHEDULER (BOOTS): SYSTEM OVERVIEW

6. PERFORMING ORG. REPORT NUMBER

7. AUTHOR(., S. CONTRACT OR GRANT NUMBER(e)

F. Laurence Keeler and Richard L. Church

9 PERFOPMING ORGANIZATION NAME AND ADDRES S 10. PROGRAM ELEMENT PROJECT. TASK

Training Analysis and Evaluation Group AREA & WORK UNIT NUMDERS

Department of the NavyOrlando, FL 32813i. CONTROLLING OFFICE NAME AND ADDRESS 12. REFORT DATE

September 198313. NUMBER OF PAGES

3814. MONITORING AGENCY NAME & ADDRESS(If different from Controlling Office) IS. SECURITY CLASS. (of thim report)

Unclassified1Sm. DECLASSIFICATiON, DOWNGRADING

SCHEDULE

16. DISTRIBUTION STATEMENT (of this Report)

Approved for public release; distribution is unlimited.

17. DISTRIBUTION STATEMENT (of the obatract entered In Block 20, If different from Report)

IS SUPPLEMENTARY NOTES

This report is one of four documents dealing with the BasicOrganizing/Optimizing Training Scheduler (BOOTS) system.

19. KEY WORDS (Continue on reverse side if neeceeary and identlify by block number)

Accession Training Scheduling RecruitsTraining Computerized SchedulingRecruit Training Command Scheduling ConstraintsRecruit Training Curriculum Scheduling Algorithms

Lambda-Opt20. Al.TS ACT (Con nue on revere. ci e If neceey "d Ian dentify , bck m ^ be,.

.__iThis report provides an overview of the 'asic Organizing/OptimizingTraining Scheduler (BOOTS) system and its operation. The BOOTS is acomputerized tool designed to aid in the creation of Master TrainingSchedules for each Navy Recruit Training Command. The system is defined interms of three operating functions: _A data file editing in which the ,userdefines the scheduling parameters and constraints, O. scheduling, in whichthe user either manually builds the training schedule while using the--

(continued on reverse)FORM

DD I J AN73 1473 EDITION OF I NOV65 1sOSOLETE UnclassifiedS N 0102- IF- 014- 6601 SECURITY CLASSIFICATION OF THIS PAGE (When Date Entered)

----------------------- : ".

Unclassified\ SECURITY CLASSIFICATION OF THIS PAGIE MM DOS mltIl

SC " 20. ABSTRACT (continued)

system to monitor all of the defined constraints or directs the system to

optimize the schedule in terms of these constraints, using a Lambda-Opt

heuristic, and (a) data output generators for formatting and printing hard

copies of the resultant schedules and supporting data.

SN 0102- LF- 014- 6601Unclassified

SECURITY CLASSlPICATIO0 OF TillS PAGE(lhen Doea E uteEl)


ACKNOWLEDGMENT

Appreciation is tendered to CAPT B. R. Nyce, CAPT P. B. Boyne, and CAPTJ. K. Gardella, commanding officers of the three Recruit Training Commands,for their support. Their staff personnel provided much needed expertiseconcerning the recruit training process. CAPT L. W. Fernald (USN retired),former Commanding Officer RTC Orlando, is also cited for his unfailing encour-agement and support during the early stages of the program.

The many contributions of the following personnel are gratefullyacknowledged: LT T. J. Hoskins, MMCS R. R. Ward (retired) and MSCSA. A. Gray, RTC Orlando; LT J. A. Somers, LT P. A. Hagan, CW02 W. 0. Brown,and OSCS J. L. Lockwood, RTC San Diego; and LTD. L. Hunt and LT C. R. King,RTC Great Lakes.

The following TAEG professionals have contributed substantially to thesuccess of the program. Mr. D. R. Copeland initiated the effort to developa computerized scheduling system for the RTCs and served as the team leaderfor the project. Mr. M. G. Middleton provided key professional guidance inthe selection of the equipment upon which the system was implemented.

A[* . J


TABLE OF CONTENTS

Section Page

I INTRODUCTION .............................................. 3

Purpose .................................................... 4Method ..................................................... 4Orqanization of This Report ................................ 5

II THE SCHEDULING PROBLEM ..................................... 6

Associated Scheduling Functions at the RTCs ................ 7Scheduling Problem Constraints and Objectives .............. 8The BOOTS Approach to Generating an MTS .................... 8

III SOLUTION APPROACHES ........................................ 17

Alternative Approaches ..................................... 17Developing a Heuristic Approach ............................ 18

IV THE BASIC ORGANIZING/OPTIMIZING TRAININGSCHEDULER (BOOTS) SYSTEM ................................... 21

Hardware Requirements for Operating BOOTS .................. 21BOOTS System Architecture .................................. 21Data File Organization ..................................... 23Functional Programs ........................................ 25Operation of the BOOTS System .............................. 26

V SUMMARY AND RECOMMENDATIONS ................................ 28

Summary .................................................... 28Recommendations ............................................ 28

APPENDIX A Variables, Terms, and Symbols Used inMathematical Expressions ................................. 30

APPENDIX B BOOTS System Hardware Requirements ....................... 34

APPENDIX C Sample Pages From BOOTS Data Files ....................... 35

GLOSSARY OF TERMS AND ABBREVIATIONS .................................. 47


LIST OF ILLUSTRATIONS

Figure Page

I BOOTS System Architecture ................................... 22

2 Flow Chart for Creating a Schedule .......................... 27

LIST OF TABLES

Table Page

1 Major Factors Used in Evaluating a Schedule ................. 9

2 Master Training Schedule Model Constraints .................. 12

3 Master Training Schedule Model Objectives ................... 14

4 Composite Measure of Effectiveness .......................... 16

5 BOOTS Data Files ............................................ 24

2


SECTION I

INTRODUCTION

The Navy currently trains between 90,000 and 100,000 recruits eachyear. This training is conducted at the three Recruit Training Commands(RTCs) located at Great Lakes, Orlando, and San Diego. The large throughputrequires that new companies (training units) be formed and commence traininqalmost daily at each RTC. Consequently, a complex set of company trainingschedules called the Master Training Schedule (MTS) is required for eachRTC. Each company schedule must account for the entire training curriculumand, at the same time, avoid conflicts in resource utilization with anyother company schedule within the MTS. Because of the complexity of havinqmultiple companies utilizing one set of resources while operating ondifferent schedules, the preparation of the MTS is a highly labor intensivetask. The manual development of the MTS also results in errors or conflictsin the schedule. As a result, each RTC spends considerable effort ir.checking and rechecking the MTS to minimize such problems and to ensure thatadequate resources are available.

The current training curriculum1 contains 93 individual traininqfunctions and topics which must be scheduled within a total of 380 trainingperiods over a length of 7.7 weeks. Each schedule in the MTS must adhere toa set of rules which dictates the separation and grouping of these 93traininq elements. Further, special functions (e.q., Pass-In-Review) mustbe scheduled at prespecified times. Finally, the MTS must maintain aprescribed order for certain traininq events (e.q., one medical exam mustprecede TB skin test reading). Multiple companies are started almost daily(each following~a different schedule), and the MTS must ensure that facilityand staff limitations are not exceeded in any period of any day.

The RTCs experience a seasonal peak in recruit input in the earlysummer with much lower input during the winter. In addition, there areubstantial variations in the daily recruit input. Thus, the MTS must also

be able to accommodate the seasonal and daily fluctuations in the number ofrecruits undergoing training.

The scheduling of recruit training has been a continuing concern of theChief of Naval Education and Training (CNET). 2 Accordingly, in December1979, the TAEG was tasked to conduct a series of studies concerned withrecruit training. 3 This tasking included the definition of an optimumschedule for recruit training. The start date for this effort was October1981. However, a working conference held at the Chief of Naval Technical

ICurriculum Outline for U.S. Navy Recruit Training, X777-7770, RevisedDecember 199,7,T e"of Nava-echnical Training, NAS Memphis (75),Millington, TN 38054.

2The lenqth of recruit training has been changed 26 times since 1944. Thesechanges have necessitated major revisions to the MTS at each of the RTCs.

3CNET ltr Code N-53 of 6 December 1979.

ro l iiln ii ll n ui _ .. I . ..[ .:3 .


Training (CNTECHTRA) on 3-5 March 19814 resulted in major changes to thecurriculum, and the CNET priority on the development of an optimum schedulewas changed. As a result, the TAEG effort addressing optimum scheduling wasinitiated in April 1981.

In response to this tasking, the TAEG has developed a computerizedsystem designed to considerably simplify the scheduling task. A flexibleuser-oriented system, called Basic Organizing/Optimizing Training Scheduler(BOOTS) was developed with the support of the three RTCs and is currently inuse to develop a Mobilization Schedule (under conditions of nationalemergency) for RTC Orlando. An operational system for use by the schedulingofficer is being installed at RTC Orlando.

Two documents have been prepared which describe the system and itsemployment. The present report is devoted to an overview of the BOOTSsystem and chronicles its development. The second report 5 is the BOOTS UserGuide which provides step-by-step instructions on how to use BOOTS inpreparing Master Training Schedules.

PURPOSE

This report presents an overview of the concepts and approaches used inthe development of the BOOTS system. In developing the BOOTS system, themajor objective was to design a flexible, user-oriented system capable ofgenerating an MTS. This tool would simultaneously address all schedulingconstraints in order to produce an optimal MTS. In addition, the system wasto be designed so that it could be implemented at each of the RTCs and beoperated by current RTC scheduling personnel.

METHOD

The BOOTS system was developed with the cooperation and support of thethree RTCs. The RTCs' staff played an indispensable role in defining thescheduling problem presented in this report. Representatives of eachdepartment and training division were interviewed concerning how they usedthe training schedule to accomplish their assigned functions. One companyat RTC Orlando was monitored for the entire period of training including in-processing and graduation. This provided information on how the schedule isused daily by the company commander and what problems are encountered bycompanies in utilizing the training schedule. Considerable time was spentwith the scheduling officer and his staff at each RTC to gain an in-depthunderstanding of the total scheduling process. This information was used indefinirg the size and scope of the scheduling problem. A set of genericconstraints comprehensive enough to describe all of the conditions whichapply to the scheduling process was then defined.

4CNTECHTRA msg 062235Z Mar 81, Subj: Recruit training.5Richard L. Church and F. Laurence Keeler. The Basic Organizino/OptimizingTraining Scheduler (BOOTS): User's Guide. Technical Report 151, September1983. Training Analysis and ETvation Group, Orlando, FL 32813.

4


Using well known operations research methodology, a formalized inteqer-linear programming (ILP) model of the scheduling problem was then developed.Various approaches to solving the ILP model were considered and the two mostpromising approaches were selected. These were the Interactive ModeProcessor (IMP) and the Heuristic Optimizing Processor (HOP). Finally, acomputerized system of data files and computer programs was developed forimplementing the selected approaches.

ORGANIZATION OF THIS REPORT

In addition to this introduction, the report contains four sections andthree appendices. Section II describes the recruit scheduling problem. Anumber of alternative ways to solve the described problem are discussed insection III. Section III also provides detailed descriptions of the twoapproaches chosen for development. BOOTS, the computerized system that hasbeen designed to optimize the scheduling of training functions, is describedin section IV. Section V contains a summary and recommendations for use andextension of the BOOTS system. Definitions of all mathematical variablesand terms used in the presentation of the MTS scheduling problem areprovided in appendix A. Appendix B presents equipment options required tosupport the operation of the BOOTS system. Finally, sample pages fromlistings of the data files created by the BOOTS system are contained inappendix C.

5


SECTION II

THE SCHEIULING PROBLEN

All elements of the recruit training process must be scheduled to avoidconflicts of space and time. The elements to be scheduled include:

• training topics (instructional elements; e.g., fire fighting)

" training functions (noninstructional elements; e.q., medicalexaminations)

" meals (breakfast, lunch, and dinner).

The training topics and training functions, along with the number oftraining periods required for each, are specified in the Curriculum Outlinefor U.S. Navy Recruit Traininj, X777-7770, published by CNTECHTRA. Aschedule is required to specify the exact sequence and times for allactivities including meals. Meal serving times must be set in cooperationwith the galley at each of the RTCs since the galleys are operated under theNaval Administrative Command.

Although many tasks are performed in the course of operating an RTC,the development of the complex MTS is the most important and complex. EachRTC develops an MTS to accommodate its specific resources and environmentalconditions. Since the individual base resources vary, the schedules areoften quite different among RTCs. In addition, special schedules are usedat some locations but not at others. For example, RTC Orlando utilizesrecruits who are members of regular training companies to form a flag,drill, and chorus group for Pass-In-Review ceremonies. These recruits maymiss certain training events when they practice and perform these specialfunctions. The classes and exercises which are missed must be made upoutside of normal training, either during Ships Service Work Week or afternormal training hours. However, at RTC San Diego and RTC Great Lakes,special companies for flag and drill teams and a baiod (plus staff and chorusfor Great Lakes) are started each week. These companies utilize specialtraining schedules which incorporate the extra duties in lieu of the tasksnormally assigned during the Ships Service Work Week.

The major problem in setting up a schedule is the assignment of space.While space assignment may not, at first, seem difficult, the complexitiesof operating a facility where new companies begin training almost every daypresent many problems. At RTC Orlando a maximum of two companies starttraining each day, one using the Alpha schedule and the other the Bravoschedule. Given the current 38 day (7.7 week) curriculum, and two schedules(Alpha and Bravo), there is the potential requirement for 76 different dailyschedules.

6

6 The special staff and drill company schedules add to the complexity of

schedules at the other two RTCs.

6


In addition, every company starting training in the same week rvust

Pass-ln-Review on the same day (the last Friday of training). Thus, if Icompany starts on a Monday, it will Pass-In-Review on the 35th day oftraining (called the 7-5 day because it is the 5th day of the 7th week). Acompany starting on a Tuesday will have its last Friday of training and

Pas,-Tn-Review ceremonies on the 34th day of training (7-4 day). Therefor,since all companies starting in the same week fin the same traininq qrfmr))Pass-In-Review on the same day, there must be 10 (? X 5) variations of thoschedule. In scheduling, training area usage must be tallied for eachtraining area for each day of the week and period of the day for up to 76different comcanies each starting on one of 3F different days and pach usinqone of 10 variations of the 38 day schedule. Clearly, makinq up an MTS is..formidable scheduling task.

Current preparation of the MTS is a labor intensive task with few or nospecial aids being employed. Revisions made to the MTS to accommodate man-dated curriculum changes frequently fail to consider one or more schedulingconstraints (e.g., capacity, separation, precedence) and must be modified,often more than once, until all scheduling conflicts have been resolved.The scheduling process requires keeping track of a myriad of details onresource usage as well as on curriculum requirements. Computer technologyis ideally suited for increasing the efficiency of such record keeping andassignment tasks.

ASSOCIATED SCHEDULING FUNCTIONS AT THE RTCs

In addition to being used by the company commander to move his companythrough the 7.7 week curriculum, the MTS is used by every department of theRTC. The Standards, Testing and Evaluation (STE) Department uses the MTS tomake up tests. Etch test is made up according to the cestable classesassigned to the MTS before a given examination period. Furtnermore, the MTSis used by the Scheduling Officer in preparing the Recruit Plan of the Day(RPOD). The RPOD is published daily and specifies for each company intraininq, the schedule used and the current day of training in thatschedule. Any changes or deviations to the company schedules for that dayare also noted in the RPOD. The RPOD is usually made up several days inadvance.

The RPOD is used in conjunction with the MTS by all departments of theRTC. From the RPOD, a department can easily determine when and how manycompanies are scheduled for a given function on a given day. For example,the Basic Military Officer (BMO) utilizes the RPOD to schedule platforminstructors. By using the RPOD and the MTS, the BMO can estimate severaldays in advance the total number of classroom instructors needed in eachperiod. He can then schedule his platform instructors (usually on arotational basis) to meet the teaching demands. In a similar way, the STEDepartment schedules proctors for examination periods and the MilitaryInspection Department (MID) schedules inspectors. Further, the galleyutilizes the RPOD to determine the exact demand for meals and when they areto be served. Although all these scheduling tasks are largely controlled bythe MTS, they are secondary as compared to the development of the MTS and 4

are not included in the BOOTS system.


SCHEDULING PROBLEM CONSTRAINTS AND OBJECTIVES

The MTS is composed of a set of individual company schedules. Theschedules are based on the day of week that training begins as well as onspecial conditions (e.g., the requirements of special companies whichparticipate in several Pass-In-Review ceremonies). At times of peak recruitinput, all schedules in the MTS will be used simultaneously. Thesimultaneous use under peak input conditions will determine the peak demandson RTC resources and services such as classroom space, instructor pool,qalley serving capabilities, and medical evaluation teams. Each scheduleneeds to be determined with respect to the resources used by the otherschedules to prevent the assignment of more companies to an individual RTCtraining area than the capacity of that area.

In addition to capacity limitations, there are a number of otherfactors involved in assigning training events to a schedule. With theassistance of RTC personnel, 11 generic factors (including capacity) weredefined (see table 1). These factors are comprehensive and can be used todevelop and evaluate a Master Training Schedule. The 1i factors listed intable 1 form the backbone of the BOOTS system.

THE BOOTS APPROACH TO GENERATING AN MTS

Each of the factors given in table I can be defined in quantitativeterms. Together, these quantitative terms can be used to define (in aformal way) a mathematical programming problem representing the problem ofcreating an MTS. Solving the mathematical programming problem, then, isequivalent to solving the scheduling problem.

The BOOTS system is based on the definition of a formal mathematicalprogramming problem encompassing the 11 factors. The mathematicalprogramming problem is described in terms of a set of constraints (see table2) and a set of objectives (see table 3). Although a full understanding ofthese conditions is not necessary, it is important to recognize that the MTSscheduling problem has been structured using common operations researchmethodology and that this structure encompasses the 11 major factorsidentified by RTC personnel as the key factors for developing an MTS. Withthis criteria set, the BOOTS system can easily rate any MTS and look forchanges that can improve it. The rating of a schedule is based on a simpleweighted combination of the 11 factors to form a composite measure ofeffectiveness. This rating system is given in table 4. The BOOTS system isdesigned to either assist in developing an MTS (using the IMP) orindependently generate an MTS (using the HOP). In either case, the resultis an MTS complyinq with this set of factors.

The mathematical programming problem structured in tables 2 and 3 is anILP model of the MTS scheduling problem. With a minimum of 10 different

schedules, 38 days of training, 10 periods per day, and 20 differenttraining areas, over 4 million variables would be required (and almost 4million constraints) for its solution using conventional integer programmingtechniques. Because of the enormous size of the problem, it is impossibleto solve the BOOTS model with a straightforward integer programming softwarepackage. Section III of this report presents a set of alternativeapproaches to this problem and discusses the techniques selected for

implementation in the BOOTS system.

8


TABLE 1. MAJOR FACTORS USED IN EVALUATING A SCHEDULE

Factor Definition Requirement

Clustering Formulation of a To groLp and/or divide the trainingrestriction element by periods allocated to the traininggrouping one or more topics and functions designated by thetraining periods from curriculum outline into schedulableone or more training training events (e.g., the grouping oftopics/functions, stencilling with uniform issue or theand designating the dividing of the 32 periods of PTnumber of training into several separate training events).events complying withthe constraints of thisrestriction elementwhich must be assignedin the schedule

Assignment Inclusion of each To insure that every training eventtraining event in the and thus every training elementschedule and every training topic or function

in the Curriculum Outline is includedin the schedule.

Blocking Preferred scheduling To logically group compatibleof two or more training events when constraintsdifferent training do not dictate otherwiseevents together (e.g., the grouping of Company

Commander time before a scheduledbarracks inspection).

Type Dispersion of To avoid the concentration of anySmoothing similar types of one type of training on any one day

training events (e.g., the banning of more than onesession of physical training on anysingle day).

Separation Prescribed minimum To maintain prescribed periods ofand/or maximum number time between certain medicalof days between pairs functions or to establis" suffi-of training events cient time between training events

for necessary auxiliary functionsto have occurred (e.g., a 2-week

period between the preclassifica-tion briefing and classifying toallow time for all necessary paperwork to be accomplished).

9


TABLE 1. MAJOR FACTORS USED IN EVALUATING A SCHEDULE (continued)


'Conqruence Keep all variations Because all companies which start train-of a given schedule ing in the same week must Pass-in-Reviewas similar as on the same day of the final week, theirpossible schedules must be different at least

during the last week of training. How-ever, because these same companies com-pete against each other in the competi-tive inspections and testing system,their training schedules should be asidentical as possible.

Day Prescribed day-of- To maintain a required order ofiPrecedence traininq relation- precedence of training for one or

ship one training more training events (e.g., theevent must have 1st academic test must be precededto another by all of the traininq that is tested

by it,.

Period Proscription from To avoid schedulinQ certain types ofPrecedencp havinQ certain train- training events in incompatible orders

ing events precede (e.g., innoculations must not be fol-certain other lowed, on the same day, by physicaltraining events on training).the same day

Window Days/periods of To establish those days and periodstraining in which of training upon which a traininga particular event may be (and may not be) sched-training event uled (e.g., The Uniform Code ofmay be scheduled Military Justice must be scheduled

during the first five days of train-ing; or, that the issuing of uniformsmust be during those periods thatsupply is open).

Preference Relative preference To distinguish between those daysfor scheduling thp and periods on which it is highlytraining event for desirable to schedule a trainingeach day/period of event from those on which it isthe window less desirable (but still feasible)

(e.g., it would be most desirable toschedule the final performanceevaluation the last day of trainingwith decreasing desirability as thetime from that day increases; or,it would be most desirable toschedule physical training early inthe morning, as compared to later inthe day).

10


TABLE 1. MAJOR FACTORS USED IN EVALUATING A SCHEDULE (continued)


Capacity Keepinq the number To avoid exceedinq the physicalof companies using capacity of any traininq area durinqa traininq area less any period of any day of the week.than or equal to itscapacity at all times

Movement Restricting the move- To avoid having the training area ofments from one train- one training event so far from theing area to another training area of the precedingto those moves which training event that it cannot bemay be accomplished reached in the allotted time betweenwithin a prescribed periods (e.g., the schedulingtime, and to minimize of physical training on the drilltotal movement field following a barracks inspection

in San Diego).

11


TABLE 2. MASTER TRAINING SCHEDULE MODEL CONSTRAINTS

FACTOR MATHEMATICAL EXPRESSION

Aw . i lrirt E Xsed for each s and e.d pp

Capacity R erD swX p C rfor each r, w and p.*s e d er dwed r

Window Xsd 1 for each s and e.d 6D pP sede e

Periy Xsd + 1 for each s, e, d and p.

TPeo Xsd + X X <sdI ~ for eachs, e, dand p.

sedp seS~ + dps'dpPrcdec d'=1 Fe dep d'=d=

Typ Xs sed' + sfp for each s, e, d and p.Smoothin d '-d v

e

12 e ~


TABLE 2. MASTER TRAINING SCHEDULE MODEL CONSTRAINTS (continued)

VERBAL DESCRIPTION

Each traininq event, e, must be assigned once in each schedule, s, to aday of training, d, and a training period, p, within that day.

The usaqe of no training area resource, r, may exceed the capacity, Cr,(if that resource.

Each training event, e, must be assigned to a day, d, within its daywindow, De, and a period, p, within its period window, Pe.

No training event, e', from the set of training events, Ee, prohibitedfrom preceding training event e will be assigned to a day whichprecedes the day, d, to which training event e has been assigned.

No training event, e', from the set of training events, Fe, prohibitedfrom preceding training event e on the same day of training will beassigned to an earlier training period, p', on the same day oftraining, d, as that assigned to training event e.

No other training event, e', from the same smoothinq type set, Te, astraining event e will be assigned to the same day, d, as training evente.

No training event, e', which is of the set, Se, of training events tobe separated from training event e, will be assigned to a day, d',which is not within the minimum, ue and maximum, ve, range of the day,d, to which training event e has been assigned.

13


TABLE 3. MASTER TRAINING SCHEDULE MODEL OBJECTIVES

FACTOR MATHEMATICAL EXPRESSION

Day Max Ped XsedpPreference s e d p

Period Max Qep XsedpPreference s e d p

Blocking Max Z (X +s e d p e'ee

( Xsedpl + se'dp+l )

Movement* Min D, Ms d p r r rr sdprr

*The following constraints are necessary to define the movement, Msdprr,

variables.

a) X sedp Rer Msdprr , for each s, e, d and p.sedp r r

b) X E Rer M for each s, e, d and p.Xsedp r r sdp-lrr

c) L r Msdpr r , < 1 for each s, d and p.r rpr

14

L. . =I= ' dL .. . . . " '. . . . . . . IIII lUU l ln ln n n n I I I I I I I -' " '


TABLE 3. MASTER TRAINING SCHEDULE MODEL OBJECTIVES (continued)

VERBAL DESCRIPTION

Maximize the day preference, Ped, of the day, d, to which each traininqevent, e, is assigned.

Maximize the period preference, Qep, of the period, p, to which eachtraining event, e, is assigned.

Maximize the number of training events, e', within the set, Be, oftraining events to be blocked with training event e which are assignedin periods adjacent to the training peric' p, to which training evente is assigned.

Maximize the distance, Drr', between the training area resources, r andr' required by the training events, e and edp+1 assigned to adjacentperiods, p and p+1, on the same day, d.

15


TABLE 4. COMPOSITE MEASURE OF EFFECTIVENESS

CMOE =+W XASSIGN s e d p i

( E E E Ped X sedpDPREF s e d p

+ W ep XsedpPPREF s e d p

+ W } XsedWINDOW s e dee PePe

+ W Xs e d p + ' ( se'dp- + xse'dp+lBLOCK s e d e eEB e

+ +W 0 D+ I (}1 Z }1 1 1 er Orw p Dsdw Xsedp)

CAPAC s e d p r w

W dr' r s r Mr,MOVE s e d p r r

-W x) )DPREC s e d p sedp e'ZEe d'=l ' se d'pl

-W F. E1 E E ( xsep E- E xsdp)

PPREC s e d p sedp e'eFe p=1 se'dp'

-W ( E E E ( Xsedp + IE Xsedp,TYPE s e d p e'ET e P'

SEPAR s e d p sedp eSSe p '

e_ Xse'd'p, + X se' d'p,d'=dv e d' =d+u e

16


SECTION III

SOLUTION APPROACHES

The MTS scheduling model was defined in section II on the basis of 11different criteria. From these criteria, a mathematical programming problemwas developed. The mathematical programming problem given in section II canbe classified as an integer-linear programming model with over 4 millionvariables and constraints. This ILP problem is too large to be solved bythe straightforward application of well-known integer programming software.This section describes the conceptual issues involved in solving the MTSscheduling model using alternative approaches. It is not essential for theuser of the system to be familiar with these issues; however, the issuespresented here, along with the factors presented in section II, form thebasis of the BOOTS system design, which is described in section IV.

ALTERNATIVE APPROACHES

The BOOTS model is extremely large in both constraints and variables.As indicated earlier, the size of this model precludes the use of a generalpurpose mathematical programming solution technique. However, there areseveral alternatives which can be considered in generating a solution.These alternatives include:

* solve the programming problem by decomposition

* segment the problem:

assign the training events to days

arrange the training events within each day to make acomplete schedule

place in the schedule all training events associated with stronqconstraints and then optimize the remaining training eventassignments

assign optimally each training event to a week (or other period oftime) based upon windows, smoothing, and precedence, then optimizethe final placement of those training events within the week

devise a sampling/generation program to develop a schedulequickly; rerun the system many times and pick the best schedulegenerated

structure a simpler problem using preestablished rules to form thetraining events into groups, then solve the simpler problem

package by a network code

allow the scheduler to place or move training events in theschedule, while providing updates on all constraint violations andassistance in finding appropriate open training periods for

17


nonscheduled training events (The scheduler would thus interactwith the computer in developing a schedule.)

* qroup all training events on the basis of which week of trainingthey should be scheduled and then optimize the weekly schedules

" develop a heuristic procedure to formally analyze the entireproblem simultaneously.

Each of the above approaches is a feasible alternative givenappropriate computer resources and personnel. Because of computerlimitations, the first five approaches were eliminated from consideration.Approach 6 was eliminated because it was not entirely clear that the problemcould be simplified without eliminating some difficult to handle butnecessary conditions. This left three approaches. Approach 7 utilizes thecapabilities and expertise of the scheduling officer in generating aschedule in an interactive fashion. Since new schedules are oftenmodifications of existing schedules, the interactive approach could also beused to speed up such modifications while making sure that no chanaeviolated existing constraints in attempting to meet new objectives. Theimplementation of this interactive approach in the BOOTS system is calledthe Interactive Mode Processor (IMP).

An optimizing capability is also required to improve on schedulesdeveloped by the IMP or to develop a completely original schedule. Thegrouping approach (approach 8) and the heuristic approach (approach 9) areboth capable of meeting the optimization requirements and both are

compatible with the IMP. The grouping approach could accept the trainingevents grouped into the various weeks by the scheduler and optimize theschedule for each week. The heuristic approach (approach 9) could bestructured in many ways but in any case it could use the same routines asthe IMP to keep track of constraining conditions. In fact, many of the IMPcomputer routines could become part of the heuristic. It was because ofthis possibility that a heuristic method was chosen. The implementation ofthis heuristic approach is called the Heuristic Optimizing Processor (HOP).It should be emphasized that the grouping approach utilizing IMP as thefirst step is still a viable option, even though it was ruled out for theearlier developmental stages of the scheduling system. At some future time,resources and needs may dictate the further development of this procedure.

DEVELOPING A HEURISTIC APPROACH

The MTS scheduling problem to be solved falls into the wider area ofapplication known as combinatorial optimization. This area encompasses avariety of logistical problems ranging from the optimal design forcommunications and power distribution networks to the optimization oftransportation and warehousing. Scheduling is also included in this classof problem.

Many possible heuristic procedures have been developed to solvecombinatorial optimization problems. Of these, one type of heuristic hasproved to be very robust. This procedure is called the A-opt (lambda-opt)

18


method. 7 Since the A-opt method was first developed, it has been appliedto many types of combinatorial optimization problems including locationanalysis with good to excellent degrees of success. 8

A A-opt procedure can be explained using a simple location problem inwhich five warehouse locations are to be selected from 75 potentiallocations to serve a set of military bases. Assuming that a reliable costfunction is available, the problem is to find the five sites out of a totalof 75 sites which provide the least cost supply system. At first, one miqhtsuggest enumerating all possible combinations of five sites, evaluating eachconfiguration generated, and selecting the combination having the leastcost. Unfortunately, there are 17,259,390 possible combinations and itwould be impossible to solve this problem by enumeration in any reasonableamount of computer time. However, as long as the model is reasonablylinear, a A-opt technique, using relatively few computer resources, can beapplied to this problem with an extremely hiqh probability of finding theoptimum or a close to optimum solution. A description of the A-optprocedure with A=I as applied to the five facility, 75 potential siteproblem is given belong.

Start with a set of any five sites. Call these the facility positions.Pick a potential site (not a facility position) and determine if replacementof any facility position by this site would yield an improved solution. Ifone or more replacements by this site would yield an improved solution, makethe replacement which will yield the greatest improvement. Then pickanother potential site and determine if its use as a replacement for any ofthe five facility positions would yield an improvement in the solution, andagain substitute this site for the facility position where the bestimprovement can be made. Continue to select new sites and compare them withall facility positions until all potential sites have been considered once.This represents one complete cycle. Perform additional cycles until nofurther improvements occur during a cycle. This marks the end of the 1-optheuristic. In most cases the 1-opt solution is close to if not optimal forthis type of location problem. The 2-opt method is similar to the 1-optmethod but two-at-a-time replacements are considered instead of one-at-a-time replacements. In the case where the value is equal to the number offacility positions, the final solution will be globally optimal. Forexample, if a 5-opt heuristic were employed on this problem, an optimalsolution would be obtained. However, this would be the equivalent ofcomplete enumeration since there are 17,259,390 possible ways of pickingfive replacements. In short, the number of computations rises withincreasing values of A. Practically speaking, the A-opt procedure is

computationally burdensome on most problems when Ais qreater than two, andin many cases the 2-opt solution is not any better than the 1-opt solution.Therefore, in practice, #=I is most frequently employed.

7S. Lin. "Computer Solutions of the Traveling Salesman Problem." TheBell System Technical Journal, December 1965, pp 2245-2269.

8R. L. Church and C. S. ReVelle. "Maximal Covering Location Problem."

Paoers of the Regional Science Association, 1974, Vol. 33, 101-118.

19


The A-opt procedure for solving a scheduling problem may be describedas follows. Assume a schedule is to be optimized. In order to concentrateon making improvements where problems occur, pick a training event (or meal)in the schedule which creates a "problem" (e.g., violates a constraint formovement, separation, or windows). Check to see if any event (or meal)scheduled elsewhere when switched with the event under consideration willyield an improvement. 9 Many training events may not be eligible for theswitch because of constraints and restrictions. If an eligible switchyields an improvement, make that switch, then pick another training eventthat is associated with a scheduling "problem" and repeat the above process.The implementation of this heuristic procedure for optimizina the schedulewill be called the Heuristic Optimizing Processor (HOP).

The IMP and the HOP, described previously, are the two solutionapproaches selected for implementation in the BOOTS system. The systemarchitecture of the BOOTS system is described in section IV.

91n terms of the composite measure of effectiveness presented in table 4.

20


SECTION IV

THE BASIC ORGANIZING/OPTIMIZING TRAINING SCHEDULER (BOOTS) SYSTEM

This section provides an overview of the Basic Orqanizing/OptimizinQTraining Scheduler (BOOTS) system. It contains discussions of hardwarerequirements, system architecture, data file organization, file editors,processing Droqrams, output programs, and operation of the BOOTS system.

HARDWARE REQUIREMENTS FOR OPERATING BOOTS

Operating the BOOTS system requires the following hardware:

* minicomputer with at least 128k bytes of memory" disc drive for processing and storing the requisite data files" video display terminal for interacting with the computer" printer for creating hardcopies of the created files.

The BCOTS system software has been written for operation on a WANG 2200MVP or WANG 2200 LVP computer. These computers utilize a nigh speedmultiplexable virtual processor and can service many users, each runninq adifferent proqram, simultaneously. Therefore, it is not necessary for theuser of the BOOTS system to possess a comDuter but merely to have access toone as a user. Because there are a large number of WANG 2200 MVP computersin the Naval traininq community, qaining access to a computer withoutpurchasinq is a viable alternative for potential users of the BOOTS system.If access to a WANG 2200 MVP computer is obtained, the user will require aminimum amount of equipment. This includes:

" modem and telephone line for communications* video display terminal* hardcopy printer.

Appendix B lists the specifications and estimated costs of the requireditems for both options. The total estimated costs are less than $6,000 persite for the less expensive option.

BOOTS SYSTEM ARCHITECTURE

The architecture of the BOOTS system is based on three major functions:

* inpit and editinq of data files0 schedule processing* output of schedule and data files.

Figure 1 depicts the BOOTS system architecture in terms of these threefunctions.

To accomplish thc first function, the BOOTS system contains severalfile editors. The file editors are used to develop the specific data setsassociated with a particular scheduling problem. The data sets associated

21


i I-i~7

~g

-y 1-

,AVA

II

222

,22

A-


with a given RTC will differ markedly from the data sets of the other RTCssince each will contain resource and layout data that is specific to thatRTC.

The second function, schedule processing, implements the interactiveapproach and the heuristic optimizing approach discussed in section III ofthis report. These two approaches are accomplished by the IMP proqram andthe HOP program, respectively. These processor programs use the data filesdeveloped by the first function to help generate a coded version of theschedule which is then stored in the Schedule File.

The third function uses the Lineprinter Output Generator (LOG) to printcopies of the different data sets and to print the intermediate and finalversions of the schedule. With these three functions, the BOOTS can assistin generating MTSs encompassing a wide range of scheduling objectives andconstraints.

DATA FILE ORGANIZATION

The following paragraphs describe the files which are a part of theBOOTS system. These files are also listed in table 5, along with the fileinterdependencies. In addition, table 5 outlines the data elementscontained in each file and the constraints and/or preference factors forwhich that data is required. As a further aid to understanding the type ofdata contained in those files, sample pages from file listings are presentedin appendix C.

1. Curriculum File. The Curriculum File contains the topic/functionnumbers, names, and the number of periods required for each training elementin the Curriculum Outline. In addition, it indicates whether all periods ofthe training element must be part of the same training event (i.e., cluster)or may be separated for placement in different training events. As afurther aid to the scheduler, a one character code may be used to indicatewhich test in the curriculum the training element must precede.

2. Facilities File. Specific information regarding the traininqareas is required to evaluate capacity and movement factors. The FacilitiesFile contains, for each training area, the name, capacity, and measure ofdifficulty in getting to and from each of the other training areas. Sinceeach RTC has a unique physical plant, each RTC will require its own distinctFacilities File.

3. Time Format File. As previously stated, the number of days oftraining as well as the number of training elements in the curriculumoutline frequently changes. In addition, the number of major schedules inoperation varies. Currently there are two at RTC Orlando, three at RTC SanDiego, and four at RTC Great Lakes. This information along with other timerelated information is contained in the Time Format File. It containsinformation regarding the number of major schedules, the number of trainingdays, the number of training days in a week, the number of periods in a day,the name and starting and ending times for each period.

23


C 1.0 4,'0~ u-E~ ~~~ V --- C

'0" C 3.0 OCU(

a) U A 0 s- -. C C'0 L C '0. -..- '.. r- C0 -v Cfu C.' W0 C0- 1C. C -0

0C P'0 wC4j CE . 0~ 04 xC r0 .

0 ~ ' T0 4 E C 1-4.V.

0 M> 0C 44'' u .; C I0

oG 0' 0' M0 4, C"4 01 C0 C 0 0

'0M 0~ A~ v00 'V CJ'' 4 M.4 CU 0Cr

CI Cj .0 dx (U "

-- j CL UE

00 0 '0L '0LI0C4Jo '400 V .0'00L 4 -.0' W0 w- C0 00 -0 S-4'. '00-0 0'.'0"-

,x ~ ~ " rIG r- r w I ) ) WMC1 u m0 1 C'.'. >1 ~ ' LU >. x u. . ' 4 0 -- OCz

4.4 u E. '010 .' 0 ~ l 0- '0 '0 w- T, C'0 ' 0. v10xC ' W0 )m. C00~ C4. MC s- 10 V) 4) C 0 -E CO 0 4 L

a, LI O,0-r C C 31 a - O I 0C 0- S.0-r-( C u .0 u ' 0 0.Lgm .0'0 '0 0 Q M ' QC aL (U U) M '0 0.C0 C'.E C CI w

0u Im 0 0 0" '0 M 0 '0 - S - W 3 0 a.JZZw '0.-> '0 C 0'0v ' v 0 C a)L-L>,I<'.-1L C C '0

0...- W.C0 .10 0.E0 U 4I. C L -' 4- C- M 0W I.0- 0' W -Cm

C U & 2 C 0 ' E '0 > 1> 0 CU > 1 *.'.4I 0 0 0 0. 4)- -'V I 0 ' >,L

I=. 01 m' 0D 04 ' 0L 04 4) '0 C V S'0 - L.- 4 10 10 4- Q04CC '0.--0 2= ;rJ0E- wV-0 VV-V0 C-C.'O4~.0I .

0~ ~~~~~~ 1- 390 4--'4-' (U1 aU~l0 0 ' 00.00 E.-'0~ ~~~~ ~~ 0I.- uLLLL I.-' ." 4 0~001.- -.

VI C U) S- w0 Cu 4-

-) C'C L)-Z I0 L--'.' -0 C. M0. 0. a, u . .

C- - 72C L 2 , 00 0C0 '0 '0 01

IL a. Cr gm 91 C '0 V

V.4 04 01.. -C ' I

IV Cu .4 wO . C. 0. CC

a,> C E . 4, £ C 4, C . 0

0 0

CE

'0 0 0

0.. LIU

IA. - 24


4. Restrictions File. The Restrictions File contains the traininqevents defined in terms of the training elements designated in theCurriculum File. These training events are the smallest, inseparable,schedulable entities which may be manipulated by 'he IMP and the HOP. Inaddition to a short name and the number of periods required, theRestrictions File contains the other constraints applicable to each traininqevent; e.g., the traininq area required, the preference for each day andeach period of the day, and the training events it is to be hlocked with,separated from, and preceded by. Since each traininq event consists of oneor more periods from one or more training elements from a coded CurriculumFile, and must be assigned a training area, a specific Curriculum File andspecific Facilities File must be referenced by the Restrictions File.

5. Schedule File. This file designates the training event assignedfor each period of each day for each schedule. This information iscontained in the Schedule Layout Assignment Matrix (SLAM). The SLAM is thelargest array in the BOOTS system and is the coded version of the MTS. TheSchedule File also contains three auxiliary arrays for the additionalinformation that is required for schedule processing. Two cross referencinqarrays, one for training events comprised of training elements and one formeals, serve as directories as to when each training event and meal isscheduled. The third array is a facilities usage array which contains thetraining area usage for each period of each day of the week.

Because the Schedule File requires information regarding the number ofschedules, days, and periods, it must reference a specific Time Format File.It must also reference a specific Restrictions File for data concerning thetraining events to be scheduled.

FUNCTIONAL PROGRAMS

The computer programs required for accomolishina the three major BOOTSfunctions are divided into groups: file editors, schedule processinaprograms, and output programs. All programs for the BOOTS system have beenwritten in WANG BASIC-2 for operation in conjunction with a WANG 2200 MVP orWANG 2200 LVP computer because of the availability of WANG computers i' theNaval Education and Training Command. However, the software could bereadily adapted for use on other hardware.

FILE EDITORS. Each of the first four data files listed in table 5 has afile editor associated with it. The names of these file editors correspondto the associated file names and each file editor is used to input or editthe data concerning the factors controlled by that file.

SCHEDULE PROCESSING PROGRAMS. Schedule processing programs are required forthe interactive mode approach and the heuristic optimizing approachdescribed in section III. The IMP program, which is the implementation ofthe interactive approach, permits the scheduling officer to insert trainingevents into the schedule and to move training events within the schedule.It will automatically monitor each of the 11 criteria factors associatedwith each such assignment (see table 1). Thus, if queried, the IMP programcan instantly inform the user of how many more companies may be assigned a

25


particular training area at a particular time. The HOP program is theimplementation of the heuristic approach using a A-opt technique. Unlikethe IMP, the HOP will not provide a means for user intervention (other thanto prescribe a specific number of cycles), but it will independentlyimplement the A-opt technique, swapping training events a pair at-a-timeuntil either the prescribed cycles have been completed or no furtherimprovement, as reflected by the composite measure of effectiveness (seetable 4), can be made via pairwise swapping.

OUTPUT PROGRAMS. A set of programs exists for generating hardcopies of eachof the files. In addition to producing a hardcopy of the MTS, the programshave the capability of producing hardcopy extracts of the MTS, each designedto meet the needs of a particular user group; e.g., Medical/DentalDepartment, Military Inspection Department, Basic Military Officer.

OPERATION OF THE BOOTS SYSTEM

The BOOTS system is designed to be a tool to aid in the preparation ofMTSs for the RTCs. The following paragraphs describe the operations to beperformed and the order of their performance in the normal use of BOOTS.Reference to the procedural flow chart in figure 2 and the BOOTS system

architecture depicted in figure I (p. 22) will aid in understanding thisdiscussion.

It is assumed that there are no existing files, and that the user willhave to create new files using the appropriate file editor in each case. Asindicated in figure 2, the first step is to create a Curriculum File usinqthe Curriculum File Editor. Then, the user will be directed to create aFacilities File and a Time Format File, each using the corresponding fileeditor. Each of these first three files is independent, so the order inwhich they are created or edited is unimportant. The flow chart will nextdirect the user to create a Restrictions File using the Restrictions File

Editor. This file will reference the first two files since it is dependentupon data contained in them.

Finally, when the Restrictions File, along with the Curriculum,Facilities, and Time Format Files, is ready, the Schedule File may becreated. This is done using IMP, the interactive mode processor, to insertthe training events defined in the Restrictions File into the SLAM array ofthe Schedule File. The user may then use the HOP, which will apply theA-opt heuristic to optimize the schedule. Following application of theHOP, the user may make some final adjustments to the schedule by using theIMP once more. When an acceptable schedule is achieved, the LineprinterOutput Generator (LOG) is used for creating hardcopies of the schedule andschedule extracts.

26


(IIC iiII)-golC. ArE/I Dll PRITdmit noCURRICULU~M IRCUU

FILE ~FILE U31GIURICLUVAILABLE' coatO CURRICULUM FILE

OUTILINE FILE EDITOR

YES

Fiqur 2. FlAwCLrTIE FILr Frq ASChedlES

AVALABE? SIN FAILIIES27L

FIE6DIO


SECTION V

SUIIARY AND RECOIENDATIONS

SUMARY

This report has documented the development of a computerized aid forcreating and/or modifying Master Training Schedules at the Recruit TrainingCommands. The computer system, called BOOTS, is designed to be used byscheduling personnel at each of the RTCs. A companion reportlO is beingpublished concurrently with this report and contains the necessary userinformation required to operate the BOOTS system. A third report will bepublished to document the system software and will include variable andarray definitions, flow charts, and program listings.

The BOOTS system was developed with the cooperation and support of thethree RTCs. Scheduling personnel at each of the RTCs were instrumental inthe original stages of this development. The BOOTS system is now in theinitial stages of installation at RTC Orlando. At this point, the BOOTS

system is being used to generate a mobilization schedule at RTC Orlando.The marginal cost of installation at the other RTC sites has been estimatedto be less than $6,000 per site if timesharing access to a WANG 2200 MVPsystem is made available.

In addition to being used as a scheduling tool at the RTCs, the BOOTSsystem could be used as a planning tool by CNTECHTRA. Given copies of theFacilities Files created by the RTCs, CNTECHTRA could use the BOOTS systemto help evaluate the impact of proposed curriculum changes on each RTC.Because CNTECHTRA already owns WANG 2200MVP hardware, the marginal cost ofinstalling the BOOTS system at CNTECHTRA would be minimal.

The structured data files, which are an integral part of the BOOTS sys-tem, provide a repository for the substantial scheduling information that,to date, was not available from a single source. This file information canbe used in associated scheduling tasks outlined in section II. Theseinclude: developing the Recruit Plan of the Day (RPOD); schedulinginstructors, examination proctors, and inspectors; compiling tests andexaminations for course groups in flexible schedules; and scheduling galleyfunctions for smoother galley operations over all levels of fluctuatinginput.

RECO*IENOATIONS

As a result of this study and system development, the followingrecommendations are made:

. demonstrate the BOOTS system at RTC Great Lakes and RTC San Diego

1ORichard L. Church and F. Laurence Keeler. The Basic Oraanizin/OptimizipgTrininq S~hedur (BOOTS): User's gide. Technical Report 151, September1983. Training Analysis and Evaluation Group, Orlando, FL 32813.

28


* make the BOOTS system available for use on a continuing basis ateach of the three RTCs

provide, as a planning tool, the BOOTS system software toCNTECHTRA for use by recruit traininq manaqement planners onalready existing WANG 2200MVP hardware

expand the BOOTS system to handle the other more frequent sched-uling functions related to the MTS. These include:

scheduling galley functions for smoother galley operations

developing the Recruit Plan of the Day (RPOD)

scheduling personnel in the Basic Military Office (BMO);Military Inspection Department (MID); Standards, Testing andEvaluation (STE) Department; and the Medical and DentalDepartments.

29


APPENDIX A

VARIABLES, TERMS, AND SYMBOLS USED IN MATHEMATICAL EXPRESSIONS

VariablesTerms FunctionalSymbols Purpose Definition

Be Blocking The set of training events to be blocked with

training event e.

Cr Capacity The capacity of training resource r.

De Window The set of training days on which trainingevent e may be assigned.

Drr' Movement The measure of difficulty of movement fromtraining resource r to training resource r'.

Dsdw Capacity A variable indicating whether training day doccurs on day of the week w in schedule s ornot.

cI if in schedule s, training day dDsdw occurs on day of the week w;

otherwise.

Ee Day The set of training events which must notPrecedence precede training event e.

Fe Period The set of training events which must notPrecedence precede training event e on the same day of

training.

Msdprr' Movement A variable indicting whether movement betweentraining areas r and r' occurs on schedule s,day d, after period p.

(i, if movement between areas rMsdprr' and r' occurs on schedule s,

day d, after period p;O, otherwise.

Orwp Capacity A variable indicating whether the capacity Crof training resource r has been exceeded onday of the week w in eriod p or not.

Grwp s T 1 Rer Dsdw Xsedp>Cr;

Lo, otherwise.

Pe Window The set of training periods in which trainingevent e may be assigned.

Ped Day The relative preference for assigningPreference training event e on training day d.

30



Qep Period The relative preference for assigning trainingPreference event e in training period p.

Rer Capacity/ A variable indicating whether training event eMovement uses training resource r or not.

( if training event e uses trainingRer = resource r;

0, otherwise.

Se Separation The set of training events which must beseparated from training event e by a specifiedminimum and maximum number of days.

Te Type The set of training events which are of theSmoothing same smoothing type as training event e and

from which only a limited number may beassigned on any one day.

WAssign Weighting The relative importance of assigning alltraining events in each schedule.

WBlock Weighting The relative importance of assigning trainingevents which are to be blocked to adjacentperiods in the schedule.

WCapac Weighting The relative importance of assigning no moretraining events which use a particulartraining resource on any day of the week andperiod of the day than that training resourcecan accommodate.

WDprec Weighting The relative importance of assigning trainingevents which must not precede certain othertraining events so that they do not precedethem.

WDpref Weighting The relative importance of assigning trainingevents to days for which there are highpreferences for such assignment.

WMove Weighting The relative importance of assigning trainingevents so that the difficulty in moving fromone training resource to another is minimized.

Wpprec Weighting The relative importance of assigning trainingevents which must not precede certain othertraining events on the same day of training sothat they do not precede them on the same day.

31



WPpref Weighting The relative imp'rtayice of assigning trainingevents to periods for which there are highpreferences for such assignment.

WSepar Weighting The relative importance of assigning trainingevents which are to be separated from anothertraining event so that they are assigned to aday which is within the allowable range ofdays from i .

WType Weighting The relative importance of assigning ns morethan a limited number of training events ofany one smoothing type on any one day.

Xsedp Assignment A variable indicating whether in schedule s,training event e has been assigned to trainingday d and training period p.

C Domain Is an element of. ("aEA" means that a is anelement of the set A.)

Comparative Is equal to.

> Comparative Is greater than.

< Comparative Is less than or equal to.

Subtraction Subtract the following from the total.

+ Summation Add the following to the total.

Summation Sum the specified elements.

(ZyzyZ1+Z2+...+z )

d Training day The index of a training day.d = 1,2,3,...,a.

a Length of The last day of training.training

e Training event The index of a training event.e =

Training event The last training event.

edp+1 Next training The training event assigned to the periodevent following training event e.

32



p Training period The index of a training period.p=1,2,3...,p.

Length of The last training period.training day

r Training The index for the training resource.resource r = 1,2,3,...,F.

Training The last training resource.resource

s Training The index of a training schedule.schedule s = 1,2,3,..., s.

Training The last training schedule.schedule

ue Separation The minimum separation which may exist betweentraining event e and the separated trainingevent, se.

ve Separation The maximum separation which may exist betweentraining event e and the separated trainingevent, se.

33


APPENDIX B

BOOTS SYSTEM HARDWARE REQUIREMENTS

Equipment Vendor Name and Part No. Function Costl Cost2

Central Processing Wang 2200/LVP-32X Operate BOOTSUnit (CPU) with a minimum of functional

128K bytes user memory programs5M bytes disk storage $ 9,984 -

Telecommunication Wang 2227B Controller Provide inter- - $ 720Rayco-Vadic 3451 Modem connect to (700)

remotelylocated CPU

Terminal Wang 2226DE Interactive Provide user/DP Workstation computer system $ 2,112 $2,112

interface

Wang 2236MXD 4-PortTerminal Multiplexer $ 1,152 -

Printer Wang 2233 Matrix Printer Provide hard- $ 2,400 $2,400copy listingsof data filesand schedules

$15,648 $5,232(5,212)

lf access to a Wang 2200/MVP or Wang 2200/LVP with a minimum of 128K

bytes user memory and 5M bytes disk storage is not available.

2 1f access to a Wang 2200/MVP or Wang 2200/LVP with a minimum of 128K

bytes of user memory and 5M bytes of disk storage is available.

34


APPENDIX C

SAMPLE PAGES FROM BOOTS DATA FILES

C-I Sample page from a Facilities File

C-2 Sample page from a Curriculum File

C-3 Sample page from a Time Format File

C-4 Sample page from a Restriction File

C-5 Sample page from a Schedule File

35


r ties F a ..- t1I(..ne.t. Cnrt.r Yed C..e -- r tcFOPSB

I TPA114,1G tU:LDZ;,G I SG I

T 4 -- G- --- :. L

I BLUE TACKET 1 _ 1

+---------------------

I F]TAEJS TrAIL I C I

4----------------+----

i F IfLD HiJUZE i 2 1--------------------- +-

I rncIL I - I

+--------------------+-I "OLITH G;'&I4DR 1 10 j

I GRI14TER 1 10 1*--------------------+-

PIIF 4 1*--------------------+-

) )7F PSD - 1 2 )+---------=-------------

, ;- r,'DAIACZ : L200 '*--------------------+-

I tIET11CAL. DLDG--C4G--- i C ;,+---------------------

I ;,rj vI C I

I FIAPL<P A-:OPF 2.C.-r12+--------------------+-

I PAI9ABE A.-ELD 10+--------------------4-

I IF r.f14TAL I 2-- - - - -- - -4-- - - -- - - +-

I SAr.PACf~,DS+------------ :7-------4-

I DE44TAL ZLD- -245 1

+ ~ r l~I I. I ... ..

----------------------

--------------------------36 - .......

T-APEPSOPC [.,36 2


V' , ... Ia- J rr ,.. i,,, ,.. i{. • rtcCOP8S

, , , -. , iU .T,, .plit/ I l %ted' , _W

i-CTbined or.rrt i

2- - - - -- -- - - - - - - -- -- --C

1.2~~ ~ -c -li -rC . - ---,- -e- - -r- 1

* .- Cr-i n c ;irc d ' .-- U r- i op Cr, rcept% 1 2

I Zichrty I- 1 1

Driver Trainina 01 I .1

i.7 ILawy 04 Armed Conf lict II I IS

.! I Naval Customs and Courte-sies(TI 4 1

e2 Ship and Aircraft Feliani6rZation 2 C I 04 1 I

S-- --------------- --------------------------

e.3 Mission of the Navy I 3 C 1 2 6 5

SNavy History and Traditions I 4 1 C I 1 1 9-------------- ------------------------------ ----------------

c I The Soviet Navy CalernI I 1 1

2.6 Ship Crganizational Structure 1 1 1 11---------------------------------- ------------------------------- ----

2 7 Chaplain, Red Cross & Navy Relief Orient ; I I 1 13---------------------------------------------------------------------------- -------

S L.d I Leave, Liberty and Conduct Ashore I I 4 4 15-------------------------------.-----------

* 2.9 Introduction to Fleet Duties & Routines I 2 I C 1 1 7------- ---------- 1-------------------- 4------------------ --------

2.10 Navy's Accident Prevention Program 1 I 2 19------------------------------------------------------------------------------------- .

2.11 1 Basic Deck Seamanship 1 2 1 C 1 4 1 21---------.------------------------------ ------------------ ------

* .12 Small Boat_.- Lii vests, Paint.nr, & Prsvt 1 2 C 1 i 23----~- - --- ---------------------------------- --------

2.13 Basic Damage Control 1 4 1 C I 3 1 2S- -- -- - - - - - - -- -- - - - - - - - - - --- - - - - --

* 14 Career In ce, tive & 1,-d cAl BFerefits 1 I I 27

----- ,----- ---------------- -------------------- --------

- 2.i I Continuing Education I I 29S---- -------------------------------------

e.16 Adanceuert Proce-ure.2 C 4 1 31----------------------------------- ------------------------ ----

F t V in.. lr r ttior. P C 33 .-

------------------------------------------ --------

2.16 ' Enlisted Service Pecord I I I f I 3s 9

2.13 r Detachment, Transfer And PReportiny I S I I 37------------------------------------------- ----------------- --- --- ----

2.20 I Gp/Indvdl Progress Evaluation & Counslng 2F? 1 M 1 39- -. . ... . ..------------- -------------.------------------ -- ----

37

U


a.' Fr..t ..- -- ; rtcTor.i3 pal&

---- --- - ----- -onfeAI ----------- I-

+ .scit n

..... .... ... ......... .. . .. . ..... ,. v .r. .. r... . . . . . . . . . . .IF t .InIn j I 1C - 1 wr 0 w. - . - -- - - - I :,

--------------------------------------------------- +---

I )..-. o in - fc. I.. . . . .. .. . I----------------------------------------------- -------

-----------------------------------------------------

I ivAq of bulklhad 31I-------------------------------------------------4---

ofrnn.,i c-ompan3o-, piati fagular company tched..le I 1

I . i....t-nr . F. drill/F 1A9 rn pn...s .. 0-140 ) Z I

--------------------------------------------------------------------------------------- - I04,'3- , d y of" -, , - wovr'I week tnat dr &1/1Ca., company can Se I C)

af0P(0 dr~l/N j-'ane rer ecnedule 01-------------------------------------------------- +-I-

t.0 1 numher of r-hddu bs f 21

38


.. . . -. t.. . . .. . --) rtcRQS -

i I I Dov 49r1,.o Pd WInd,. Depot too

* II99ht* --------------------- ~ ,"A ,

* ,-, I Jp',- I '.,, *'9lJ'pl 99l . t T t~v, Area Idipi p93 9 I1tt Lot I l t lL otl TE(Ioin-.,' i,.ie*~~ -------------------------------- - - - -

'"4"014CI BU T9.A I-CPA;ILDI9G '01 I OP I - -3 1 04 '14

* ;<'? ' I,.? .-,'r ,.- C+,ndnt ' IPA9IJICf; P4JT1.flId 901 O} 01 9 -1 I2-4 ! 04 914 I

-- ---------------- . ----- - - -- -- -------I PA.,,.' , aan9 I T 9 4r mI)lL.DIG ;01 031 1-1 4-4 9 04 14------------------------------ --------

* *4 ' 1 .4 . .. .4G 01 1 0 1 9-5. 4-4 ,.04 . ..... ..

*4 1'.'.......TA149 .........'1. 0 4 t 9 04 - 1,4 - ---------. ..

---- ---- ---------------- - ------ - ----- - -0,' f 'L I |, ,er T~aC ~ I TPAT9JIIJ~91 R91L0 0 901 9 03 "91-I -S 04 , 14T*~ 41 AT 1c ,0 1- -- - - - - - - - - - - - - - - ----------- - - ----

9la j,_ ' -'-,'9, o thd I~ r 9 TIAIII G 29JL0199 909 0 ! I 24- I 04

, 14 9T,-- - - - -- - - - - - - - - - - - - - - - ------------ ------------- - -

0.5l'I Tu,- -. Co-,-to...& + ... TRA... -----LD-- 901 '0 1 -19-S 904 914

T... .... -- --- ........ ,-...........------------- -4. I 1; 'C Faji+r' TRAI9I9C 2UILDII 4 '01 01 0 9 -1 9 - 04 , 14 9 --------------TRM N UL14 ----- - - - -------- - - - - - - - - ------------ - -- - -

------------------ ----- -- . ........0.4 .", 4 C . F ,. 1rd ,, -. I TRAii4943 P410194 '0 04 1-1 9 2-1 O 04 9 14

Z; -- 4,- - - - - - - ----

- - 7S-- - - - - - -

1-,o h - ' RIM AIDr_.--------------------------------------------IS' 00 '0.9, -' I Bnd. 9 P494140 411D19G 01 or 09 9 : 1 9 -A 4 .4 14

----------------------------- ------------

01 .+ 7 ' .... . .. .. ...... ....d.... 9 p 1 J 9 ftI . 1 4 901 -. 01.. 9 .... 9 ..... 9 04.. 9 14. 9 .. . . . . -. . . . . . .

3' +.9 ' .0 1 + - # & L I ~ r t v C Q ., d '- I A 1 1 4 1 4 G C 9 U I L 1 1 4I G 0 1 0 - ' 7" 5 0 4 1 4 '

• . .... .. ..... ..... .. -_.. + +

*- - -- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - -- -

* 9- ' - .c. 0 F P ? ' T9 1 1i SIJIL DING 901 9 02 1 -1 1 B-5 9 04 9 14 j 9- - - - - - -

, j _ . . . ..41 4... B.. .. .. .. 01 01 - - 0

P +...9' ' 9 l 9 9 I 99 ... ... .. 9 9 .9 9 91-- - - - - - - --, I - - -; I ) J_ jl . ------ ---- - - ---------- - - - -

TPT* I- - - - - - - - - - - - - - - - - - - - - - - - -- - - - - - - ---- -

-', & L. - 9 ,y C r " n It ., T,'9- 9 11 AUIL9L : 9 01 01 7-198-5 04 914 - -

. t 0 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

99, '9 'CP, , '...Ij rIr, ,tt. T999Ir '4901 10 1-1 4-5 '04 149

- - -------------------------------------- .- - - -1-' ' ! . 0 'F A Mr. . , - m n e - r . I 7 ~ i 1 4 C F A IL III 9 '0 1 0 1 1 E - S 1 9 7- 9 0 4 1 1 4 19- - -- - -

--------------- + - . _ .-- - I --------------------

I9. TP949 .,4

1r .. ,1

41

99 ' ,l J 9 lSUlL DjlVs ja 1i 0 , 9 1,-)-I )4 -5 904 :914 9- - - - - - - -

...........---..........................

- - -- - - - - -----------------------------------------4 ,+ i ).~il{ ~llm-- J~ vrM [P( ,'[II# WIILDING 101+ 03 1 61 1 G-4 04 1 4 t

.... . .'+ 9 ............r ,-'.V1 .... 9 ... .: 9'49+I01 97"-3 9 -s 02 089 I ,

- - -- --- - - -------------------------- - - - . . . . . ..

39

.. . . .. . . ... -n' ---- - - - - - - -- --


-- -- - -- -- - -- -- - -- -- - -- - -- -- - -- -- - -- -- - -- -- - -- -" . . . . ... . .- .- ----------- -----------------~C1--------------- -

'- ', ' . I ' P ' 3 4 - ' 7.1 I 8 9 1 1 ;-- - - - - - ---------------------- -------- ---U ~ ~ ~ ~ ~ 1 'FO-6 . ' - 6 2 4 6 4 6 4 6 4 -1210 '1 . 6.4 ' 6. 4.9 19 6.- 'L17#0

1-3 '6 00-0 .7 ' 3.7 46 r .9' .9. 3.7 w-1000 ' I ,17t 2.17 6.7 ' 6.9 1I ': - ------------------------------- :- - _- --

* 1 .4 0 G ' 7.0 ' 7).2 7.2 I '7.2 7.2 '1-1220 'L-1220' I 7/.4 I 7.4 ' I)-179,0 'D-It.7 6.9VI 7. b2.I'-OE- 0 'tl 2.1" ' 7.41 7:4 7.1 ' 4. ' 3.90 8:430 1LIQ ' 6. 3.81 ' 6.8 I D-1% 11

* ' 'P- 0'. ' 3.* '. 3.3 ' 3.4 I 3.4 ,L11210 'L-1210 I 3.j 3.1 I 2.61 ' 6.9 ,6-1T-

- . - _ . . . . . . _ - _.. . - -. . . . _ : --- --. : . . . . . . . . . . . . . . . ... ... .. : ---.. -- - - - - - - - - - - - - - - - -- - . ...- -

*2- 'S0(,'0' I 37?

'8.3 '8.3' 8.3 '9.3 'L-1190'i 3.1 3.1 '6.9' 3 .1 '2 : a ;710'

* . -00I 3.7 ' 2. I . ' .4'12 .7' -21 12 1 01 7 .9 " , 7 1 ' 9. ' .0 1 MO'-;

.7 O _ ''.'60',-40' 3.7 '-.1 .1. '1 . - L--- -1210 - --.6 18-. 7.3 -. - -.: - $----

P '6 C 1 3.0 1 .2 7.4 2.20 I 9.1 'L-1220 'L-1220. 3.4 8.1 1 8.1 D .1 .. 1 ':9

b 1 , - 0 S ' 6.0 ' 6.2 2.2 ' 8.2 2.2 L-20 30 IL-1230 I 9.20 3.1 3.1 .1 '

_. . .. . ... .. . .. .. .. . ... .. . +..... . .............

' r 1'-0,0 ' 3, 3.6 320O 9.2 1 9.1 ' 9.3 'L-102 0 IL-lOSO 3.1 ' 2.26 ' 226 1

4 1 '.O- 0577 ' .2 ' 2.2 ' 3.7 3.11 a 6.4 IL-1220 'L-1 22. '7.3 7/.3 ' 3.1 3 .10- .? D-161

. .. . . .. . . ..-- - - -- -- - -- - -- -- --- - - - ----- . . . . - . . .

-41 '1o~'0 ' 3.7 4.12 4.13 4.14 4.16 ' -1220 IL-1220 3.1 4.1 4.19 I .1 ' 2.: 2'-1

------------ -------------------- - -- -

• C' 'F-, O'. 3.1 '1-1230 'L-1230 1.4 1.3 ' 2.9 ' .9 .1 L-1710

... . . .. .. .. ... -------- --.. ..- -.. . .. - - - . . .

:3 -.e f-50 ' 94 ' 94 ' 4 ' 9 . ' : 9.4 22 '2.L30 L-30 1 4.19 a .1 ' .9' 47 E015

- 4 -Ft -, ' ' 3.7/ 3.7 3.7 L-1230 -L-1230 4.1 4.1 3.9 E.9 ' '1-I iCe i. 1"", '4' CC-O' 3.7 .03 9.1 ' 3.6 ' 9.3 -'-1240 'L-1240 Il 2.20 9.3 9. '26 2 6 1 170

, , 3.9'3.9'3.8' 3,.6' 3.69 I 3.813.813.8' 2.8' %.;

:3-- -- -- -4----------------------- ----- --

S3.6 3.3 3.6 ' 3.L- ' 3.6 1 3.8 '3.8 0 3. ' .8 2 .1

S. '3. ' 3 3. '3.8' 3.8 ' 3.8 9 2 3.6 1 3.6 1 3.8 ' .8 '2.6 '

, : 3. 2: " r . : 3, -L 12 3. -12 t 93. 3 3. I 3. I 3.2 2 .0 1 -,1

.......... -- ------ --- - ------ --- ---

SF I I '3.4 3. ' 3.7 ' 3. 33 3 .1 3 12.8 '3.8 ' 7.6

40


GLOSSARY OF TERMS AND ABBREVIATIONS

PageFirst

Term/Abbreviation Definition 'Appearing

BMO Basic Military Officer ...................... 7BOOTS Basic Organizing/Optimizing Training

Scheduler ................................... 4CNET Chief of Naval Education and Training ....... 3CNTECHTRA Chief of Naval Technical Training ........... 4HOP Heuristic Optimizing Processor .............. 3ILP Integer Linear Program ...................... 5IMP Interactive Mode Program .................... 5LOG Lineprinter Output Generator ................ 23MID Military Inspection Department .............. 7MOE Measure of Effectiveness .................... 16

MTS Master Training Schedule .................... 3RPOD Recruit Plan of the Day ..................... 7RTC Recruit Training Command .................... 3SLAM Schedule Layout Assignment Matrix ........... 25STE Standards, Testing and Evaluation

Department ................................ . 7TAEG Training Analysis and Evaluation Group ...... 3A-opt Lambda-opt (heuristic method for

optimization using a pair-swappingprocedure where A indicates thenumber of pairs to be swapped at a time) .... 18

Training Element a Training Function or Training Topic as listed in theCurriculum Outline

Training Event a group or cluster of Training Elements which must bescheduled together at one time

Training Function a noninstructional element of the curriculum; e.g.medical exams

Training Topic an instructional element of the curriculum; e.g. firstaid

41


DISTRIBUTION LIST

Navy

OASN (M&RA)CNO (OP-115, OP-il, OP-123, OP-112T-1)ONR (442, 270)CNM (MAT-072?)CNET (01, OOA, N-2, N-3)CNAVRES (0?)CNTECHTRA (016, N-7, N-721)COMTRALANT (2 copies)COMTRAPAC (2 copies)CO NAVPERSRANOCEN (Library)Superintendent NAVPGSCOL (2124, 32)Superintendent Naval Academy Annapolis (Chairman, Behavioral Science Dept.)CO FLECOt'BATRACENPACCO FLECOMBATRACENLANTCO NAMTRAGRUCO NAVTECHTRACEN Corry Station (101B, 3330, Cryptologic Training Department)CO NAVTRAEQUJPCEN (TIC)Center for Naval AnalysesCO FLEASWTRACENPACCO FLEASWTRACENLANTCO NAVSUBSCOL NLON (Code 0110)CO NAVTECHTRACEN MeridianDIR NAVEDTRAPRODEVCENDET MeridianCISO, SERVSCOLCOM Great LakesCISO, NTTC MeridianCO FLETRACEN, MayportCO FLETRACEN, San DiegoCO FLETRACEN, NorfolkCO NAVCRUITRACOM Orlando (2 copies)CO NAVCRUITRACOM Great Lakes (2 copies)CO NAVCRUITRACOM San Diego (2 copies)CO NAVAVSCOLCOM (0212)QIC NODAC (2 copies)CON NETC (30)

Air Force

Headquarters, Air Training Command (XPTD, XPT1A) Randolph Air Force BaseAir Force Flman Resources Laboratory, Brooks Air Force BaseCommander, tdasic Military Training School, Lackland Air Force BaseAir Force Office of Scientific Research/NL

(Page 1 of 2)


DISTRIBUTION LIST (continued)

Army

CG TRADOC (Technical Library, ATTG-1)ARI (Reference Service)CDR USATC Fort DixCDR USATC Fort JacksonCDR USATCE Fort Leonard WoodCOMDT ARMOR SCH Fort KnoxCDR MPSCH TNGCEN Fort McClellanCDR SIGCEN Fort GordonCOMDT ADSCH Fort BlissCOMDT FASCH Fort SillCOMDT INFSCH Fort BenningCDR ORGEFFTNGCEN Fort Ord

Coast Guard

Commandant, Coast Guard Headquarters (G-P-I/2/42, GRT/54)CO USCG Training Center, Cape May

Marine Corps

CMC (T, TAP-40)CGMCDECCO MARCORCOMMELECSCOLCG MCRD, Parris IslandCG MCRD, San Diego

Information Exchanges

DTIC (12 copies)DLSIEERIC Processing and Reference Facility, Bethesda, MD (2 copies)

(Page 2 of 2)

AD-AI35 103 THE BASIC ORGANIZING/OPTIMIZING ...20. Al.TS ACT (Con nue on revere. ci e If neceey "d...

Documents

Transcript of AD-AI35 103 THE BASIC ORGANIZING/OPTIMIZING ...20. Al.TS ACT (Con nue on revere. ci e If neceey "d...