Charles Herring Talk

7/30/2019 Charles Herring Talk

1/38

Viable Softwarethe

Intelligent Control Paradigmfor

Adaptable and Adaptive Systems

[email protected]

http://www.charles-herring.com


2/38

Overview

Motivation and Approach Systems Thinking: MVC Revisited

The Viable Software Architecture Generative Programming

Case Studies

Conclusion


3/38

Trends and Forces

Impersonal Computing 1 computer, N persons Human Out-of-the-loop

Personal Computing 1 computer, 1 person Human In-the-loop

Proactive Computing

N computers, 1 person M computers, N persons Human Above-the-loop


4/38

Direction of Evolution

The Economy

The Brain

The Enterprise

Stockholding

Conditioned Reflexes

Industrial Probability

Penny Tossing

Brownian Motion

Statistical QualityCont.

Empty

Digital Computers

Planetary Systems

Automation

Window Catch

Billiards

Shop Lay-Out

Simple Complex ExceedinglyComplex

Deterministic

Probabilistic


5/38

Approach

How do we build these types of softwaresystems? Is there any knowledge from other fields

(outside of computer science) that can help?

I chose the Viable System Model Developed by Stafford Beer Based on Information Theory and Cybernetics Claims to capture the invariants of viable

systems Viable: able to maintain a separate existence.


6/38

Model

coreDatasetOfObservers

attach(Observers)detach(Observers)notify

getDataservice

View

myModelmyController

initialize(Model)makeControlleractivatedisplayupdate

Controller

myModelmyView

initialize(Model,View)

handleEventupdate

Observer

updatecall update

attachgetData

attachcall service

manipulatedisplay

create

Model View Controller


7/38

Controller

Model

ModelcoreDatasetOfObservers

attach(Observers)detach(Observers)

notify

getData

service

View

myProcessmyActuator

initialize(Process)makeActuator

activateoutput

update

Controller

myProcessmySensor

initialize(Process,Sensor)

handleEventupdate

call update

attach

getData

attachcall service

manipulateoutput

create

feedback

commands

Systems Thinking

Actuator

Sensor

Process

Environment


8/38

Viable Software

A viable software system is Adaptable

over time to be Adaptive.


9/38

Separate Control

/ Environment:

/ Process: *

disturbances

responses

command

control

/ Controller:

Process Model

Control Rule

Computation

Present

Critical Variables

results

command


10/38

Regulator

/ Environment:

/ Process: *

disturbances

responses

command

control/ Regulator:

response

responseplans

*

plans

/ Controller:

Process Model

Control Rule

Computation

PresentPlan

Critical Variables

results

command


11/38

Auditor

/ Environment:

/ Process: *

disturbances

responses

command

control/ Auditor: / Regulator:

audit reports

audit requests

audit reports

audit request

response

responseplans

**

plans

/ Controller:

Process Model

Control Rule

Computation

PresentPlan

Critical Variables

Queries

results

command


12/38

/ Environment:

/ Process: *

feedback

disturbances

responses

command


audit reports

audit requests

audit reports

audit request

response

responseplans

/ Adaptive Controller:

**

controlrules

observations

plans

/ Controller:

Process Model

Control Rule

Computation

Self Model

Rule Base

Present

Future

Plan

Critical Variables

Queries

results

command rules

Adaptive


13/38

/ Environment:

/ Process:

policy

*

feedback

disturbances

responses

command


goals

results

audit reports

audit requests

audit reports

audit request

response

responseplans

/ Adaptive Controller:

**

controlrules

observations

/ Supervisor:

controlrules

alerts

plans

/ Controller:

Process Model

Control Rule

Computation

Self Model

Rule Base

Present

Future

Actor Interface

Policy Base

Plan

Critical Variables

Queries

Supervisory

Viable

SoftwareArchitecture


14/38

Multi Agent Systems

Human

SkillsIntelligent Adaptive Systems

Applications

Control

Adapt

Supervise

Operating Systems

SoftwareSystem

Viable Systems

Adaptive

Adaptable

Control Adapt Supervise

Migration of Control SkillsThe Gradient of Software Evolution

A viable software system is Adaptableover timeto be Adaptive.


15/38

/ Regulator:

/ Controller:

policy

*

command

control/ Auditor:

goals

results

audit reports

audit requests

audit reports

audit request

response

plans

response

plans

/ AdaptiveController:

**

controlrules

/ Supervisor:

controlrules

alerts

*

/ Future:

/ Present:

/ Past:

Adaptive-Adaptive

Supervisor-Supervisor

Recursion


16/38

21

3

4

5

6

7

8

9

3

4

5

6

8

9

8

0

Viable

SoftwareFramework


17/38

23

4

5

6

7

9

8

0

>

1

Viable Component


18/38

System Sub-System

6 Alert Conditions

5 Goals and Policies

4 Models, Simulations, and Data

3 Resources and Accountability Requirements

2 Initial Plans and Activities

7 Links to Other Processes

1 Initial Audit and Inspection Queries

8 Environment to Adaptive Controller Link

9 Environment to Process Link

2

3

4

5

6

7

9

8

0

>

1

ComponentTransferProtocol


19/38

Meta-level (Complex Adaptive Systems in General)

Domain-Specific level (B2B, Agent Systems, Smart Environments, )

ViableSystemModel

ViableSoftware

Architecture

DomainReferenceModel &

Requirements

DomainReference

Architecture

ApplicationSpecific

Requirements

ViableSoftware

Framework

DomainFramework

&Components

ApplicationComponent

ConfigurationTools

ComponentTransferProtocol

Generative ProgrammingProduct Line Architecture


20/38

Case Studies

Groove: Diagnosing the System Mapping Groove into the VSA Adding Adaptive Controller for UI

Smart Environments

Design and Usage Scenario Lecture Room Camera Controller

e-Commerce Evolving a B2B System B2B Contracts


21/38

Groove

Mediated Model-View-Controller Groove Levels

Groove Net and Web Services

Transceiver Shared Spaces

Tools Sets

Tool: Storage and Controlling Code

Add Adaptive User Interface Controller


22/38


23/38

U(t)

1Never Infrequent Frequent Always

V(t)

m

0 0.5-1.0 -0.5 0.0

Average

1.0

d U(t)dt

1 VerySlow Slow Fast VeryFast

0 100-200 -100 0

Zero

200

1

Invisible Fading Promoted

0 4010 20 30

Normal

50 60

Bold

Fuzzy Set Membership Functions


24/38

If Usage is Average and Change in Usage is Zero Then Visibility is Normal

Average

-0.5 0.0 0.5

Zero

-100 0 100

Normal

20 30 40

If Usage is Average and Change in Usage is Fast Then Visibility is Bold

Fast Bold

30 40 500 100 200

40mcrisp = 35

20 30

-0.5 0.0 0.5

Average

50100

Inference


25/38

InvisibleInvisibleFadingNormalBold

InvisibleInvisibleInvisibleFadingNormal

InvisibleFadingNormalBoldPromoted

FadingNormalBoldPromotedPromoted

NormalBoldPromotedPromotedPromoted

Change in Usage: dU/dt

VerySlowSlowZeroFastVeryFast

Infrequent

Never

Average

Frequent

Always

Usage:

U(t)

Visibility

V

Rule Table


26/38

Smart Lecture Room

Smart Lecture Room

C1LT T

TRoomController C3C2

T

Door

Tracking Beacon

Tracking Antenna Tracking Antenna

Tracking AntennaLecturer Camera Audience Camera

Surveillance CameraLock

SmartB

oard

Wireless Base Station

W

Wireless Base Station

W


27/38

Smart Camera

SmartCamController

SmartCamOperations

SmartCamAuditor

SmartCamPlanning

SmartCamExecutive

SmartCamRegulator

SmartCamMSMQBus

SmartCamAccountant SmartCamPlant

SmartCamMSMQBus

Camera

Controller

ImageProcessor


28/38

Startup Protocol

SmartRoom Executive Planning Operations Regulator Plant

SCExecPolicySOP

SCExecResources

Goals

SCExecResources

SCExecMsgStart

SCExecPolicySOP

SCOpsMsgCommand

SCPlantMsgModel

SCExecPolicySOP

SCExecResources

SCExecMsgStart

SCExecMsgStart

SCOpsStartParams

SCPlanPlanParams

SCOpsMsgStartCmd

SCOpsPlan

SCOpsPlan

SCOpsMsgStartCmd


29/38

Operating

SmartRoom Executive Planning Operations Regulator Plant

Event

SCOpsMsgCommand

SCPlanMsgAdvice

SCOpsPlan

SCOpsPlan

SCPlantRptRoutine (5)

SCOpsRptRoutine

SCPlanOpsUpdate

SCOpsRptRoutine

SCPlantStateChng

SCOpsStateChange

SCOpsStateChangeSCOpsStateChange


30/38

e-CommerceRetailer.com

B2B and B2C B2B Subsystems Wholesalers Shippers

Exchange Web Site B2B Controller: uses Visual Rules Studio

Executive: Contracts Planning: Operations and Market Analysis Operations: Purchasing, Accounting, Shipping,

and Contracting


31/38


32/38


33/38

Viable Software

A viable software system is Adaptableover time to be Adaptive.


34/38

Multi Agent Systems

Human

Skills

Intelligent Adaptive Systems

Applications

Control

Adapt

Supervise

Operating Systems

SoftwareSystem

Viable Systems

Adaptive

Adaptable

Control Adapt Supervise

Migration of Control SkillsThe Gradient of Software Evolution


35/38

The

ViableSystem

Legal

Financial

National

International

Technological(Software)

Customers

Corporate

Dimensions of Viability


36/38

Related Work

Mary Shaw - Control Paradigm Kokar and Eracar - Self-Controlling Software

Adaptive Object Model - Foote, Johnson and

Yoder Adaptive Strategy- Olivier Aubert

Beck, Highsmith - XP, Adaptive Software

Development Oriezy - Self-Adaptive Software


37/38

Conclusion

VSM Isomorphic to Intelligent Controller Pattern Language

Theory of Autonomy

Conways Law


38/38

Questions?

[email protected]://www.charles-herring.com

Charles Herring Talk

Documents

Transcript of Charles Herring Talk