Semantic Cities @ AAAI 2012 July 23rd, 2012 University of Tabriz Faculty of Electrical and Computer...
Transcript of Semantic Cities @ AAAI 2012 July 23rd, 2012 University of Tabriz Faculty of Electrical and Computer...
Semantic Cities @ AAAI 2012
July 23rd, 2012
University of TabrizFaculty of Electrical and Computer Engineering
In The Name of God
A new method for Conflict Detection and Resolution in
Air Traffic Management
By:
Farnaz Derakhshan
Hojjat Emami
Contents
2-28
References7
Conflict Detection and Resolution Process
2
Imperialist Competitive Algorithm (ICA)4
Content
Introduction 1
Semantic Cities @ AAAI 2012
Our Proposed Model5
Test Results & Conclusion6
Graph Coloring Problem (GCP)3
Introduction (Air Traffic Control)
Air traffic:
“Aircraft operating in the air or on an airport surface,
exclusive of loading ramps and parking areas”
(Federal Aviation Regulations and Aeronautical Information Manual, 2010 edition)
Air Traffic Control:
A service operated by appropriate authority to promote the
safe, orderly, and expeditious flow of air traffic
(Federal Aviation Regulations and Aeronautical Information Manual, 2010 edition)
Introduction Semantic Cities @ AAAI 2012 3-28
Introduction (Air Traffic Control)
Air traffic control is a complicated task, involving multiple and
dynamic controls and have high level of granularity.
having a reliable, safe and efficient air traffic management is a
fundamental and critical need in aviation industry.
Introduction Semantic Cities @ AAAI 2012 4-28
Introduction (Free Flight)
Free flight is a new concept presented potentially to solve
problems in the current air traffic management system.
Free flight method has many advantages (such as less fuel
consumption, minimum delays, reduction of the workload of
the air traffic control centers, high efficiency, and has
distributed nature)
The most notable problem in free flight method is the
occurrence of conflicts between different aircrafts’
flights.
Introduction 5-28Semantic Cities @ AAAI 2012
Introduction
One of the fundamental challenges in the current air traffic
management and especially in the free flight is detection and
resolution of conflicts.
We presented a new model for conflict detection and
resolution between aircrafts in airspace using graph coloring
problem.
We mapped the conflict detection and resolution problem to
graph coloring problem.
To solve graph coloring problem we used imperialist
competitive algorithm.
Introduction 6-28Semantic Cities @ AAAI 2012
Conflict Detection and Resolution Process
Conflict:
Conflict is the event in which two or more than two aircrafts
experience a loss of minimum separation from each other
Conflict Detection:
The process of deciding when conflict (conflict between
aircrafts) will occur
Conflict Resolution:
specifying what action and how should be to resolve
conflicts
Conflict Detection and Resolution Process
7-28Semantic Cities @ AAAI 2012
Conflict Detection and Resolution (General view)
Conflict Detection and Resolution (General view)
8-28
No
Yes
Conflict Resolution
Conflict Detected?
Environment (airspace)
Conflict Detection
Operating area- Traffic Information – weather
conditions and …
Operator or Agent
Semantic Cities @ AAAI 2012
Figure.1: Conflict Detection and Resolution (General view)
Graph Coloring Problem (GCP)
GCP is an optimization problem which finds an optimal coloring for a given graph G.
(Jensen and Toft 1995)
GCP is one of the NP-hard problems
GCP is a practical method of representing many real world
problems including:
Time scheduling
Frequency assignment
Register allocation
Circuit board testing
…Graph Coloring Problem (GCP) 9-28Semantic Cities @ AAAI 2012
Graph Coloring Problem (Cont)
Given an undirected graph G=(V, E) with a set of vertices V
and a set of edges E;
A K-coloring of G includes assigning a color to each vertex
of V, such that neighboring vertices have different colors
(labels).
GCP implemented by using a conflict minimization
algorithm
Graph Coloring Problem (Cont) 10-28Semantic Cities @ AAAI 2012
1 2
3 4
1 2
3 4
An Undirected Graph before Coloring|V| = 4|E| = 4
Chromatic number = 2
|K| = 2 (Colors : Red, Green)
Colored Graph
Figure.2: an example of coloring a graph
Search Strategies
Search Strategies
Complete
Deterministic
Evolutionary Programming
Evolutionary Strategy
Genetic Algorithms
Genetic Programming
Heuristic(Non-Deterministic)
Estimation of Distribution Algorithm
Evolutionary Algorithms
Population-based
Search Strategies
Imperialist Competitive Algorithm
Point-based
11-28Semantic Cities @ AAAI 2012
Imperialist Competitive Algorithm (ICA)
Imperialist Competitive Algorithm (ICA) is Novel Socio-politically
Motivated Optimization Strategy.
Proposed by Atashpaz-Gargari and Lucas, 2007.
is inspired by sociopolitical process of imperialism !!
since in late inception, it has been used in many
applications.
has shown good convergence and global minimum
achievement.
has a lot to do with.
Introduction(ICA) 12-28Semantic Cities @ AAAI 2012
A Big Picture of ICA
A Big Picture
Figure.3: A big picture of ICA (Gargari and Lucas, 2007)
13-28Semantic Cities @ AAAI 2012
Flowchart of the ICA
Flowchart of the ICA
start
Is there an empire With no colonies
Yes
Done
Compute the total cost of all empires
No
Nois there
a colony in an empirewhich has lower cost
than that of imperialist
Exchange the positions of that Imperialist & colony
Initialize the Empires
pick the weakest colony from the weakest empire and give it to the empire that has the most
likelihood to posses it
Eliminate this empire
Yes
move the colonies toward the imperialist
Yes
stop condition satisfied
*
*
No
End
14-28Semantic Cities @ AAAI 2012
Figure.4: Flowchart of the ICA
Our Proposed Model
In our model, the main strategy is based on: "Prevention is
better than cure“
we mapped the conflict detection and resolution problem to
graph coloring problem.
we map the aircrafts congestion area to a corresponding
graph based on aircrafts condition in airspace.
imperialist competitive algorithm has shown great
performance in both convergence rate and better global
optimal achievement, therefore we used this algorithm to
solve graph coloring problem rather than other evolutionary
algorithms.
In our model, a Global approach is used for resolving the multiple conflicts.
(the entire traffic situation is examined simultaneously, and it has the high capabilities)
Our Proposed Model 15-28Semantic Cities @ AAAI 2012
Our Proposed Model
start
Solving the GCP using ICA(Conflict Resolution process)
Map the Congestion Area to a Graph(Making Adjacency Matrix)
Monitor the Environment
Project current states to future statesby using nominal propagation method
Detect the Congestion AreaCompute Distance between all Aircrafts
in Congestion Area
Define Problem ParametersAnd other Traffic Parameters
16-28
Send new Flight Paths Plan to ExistingAircrafts in Congestion Area
End
Figure.5: Block diagram of our proposed model
Semantic Cities @ AAAI 2012Our Proposed Model
Creating the Graph
here, the nodes of graph indicate aircrafts in the congestion area and every edge between two nodes indicates a probable conflicts in the future time step, if the aircrafts continue their current flight plan.
Creating the Graph 17-28Semantic Cities @ AAAI 2012
Figure.6: Graphical display of an exampleconflict detection & resolution scenario
(creating the graph)
Conflict Detection
we use a simple conflict prediction method.
we used nominal state propagation method for prediction
conflicts between aircrafts.
the current position, heading and speed of existing aircrafts
in congestion area used for mapping the current location to
the future states.
in the future states if distance between two aircraft is less than a predefined reliable distance threshold we say a conflict is going to occur.(i.e. when in future states the protected zones of two aircraft
is overlapped then system report a conflict)
here, we focused only horizontal plan dimensionConflict Detection 18-28Semantic Cities @ AAAI 2012
Conflict Resolution
we map the congestion area to a state space graph.
we replace solving the conflicts between different aircrafts
with coloring the corresponding graph.
we used imperialist competitive algorithm to solve graph
coloring problem
a colored plan for graph is a reliable solution for conflicts
problem
in the conflict resolution process, for each aircraft, we
allocate a flight path in which this aircraft will has a reliable
distance with other aircrafts and there is no risk of conflict.
In our model, we can assume each flight path as five
directional options namely: main line, deviation to right of
the main line, deviation to left of the main line, top of the
main line and bottom of the main line.
Conflict Resolution 19-28Semantic Cities @ AAAI 2012
Test Results
To evaluate our proposed model, we use random flights
model (Archibald et al. 2008)
here all aircrafts are constrained to fly at the same altitude
and at a constant speed.
Small and instantaneous heading changes for each aircraft
are the only maneuvers of resolving conflicts.
We used supposed scenarios that these samples contain 2, 3,
4, 5, 6, 8, 12, 16, 20 and 30 aircrafts in congestion area.
In air traffic control we deal with a multi objective problem.
In this paper, our goal is providing a safe, reliable and
efficient strategy for solving conflicts between aircrafts.
Test Results 20-28Semantic Cities @ AAAI 2012
Performance Metrics
The ideal state in conflict resolution model is that the aircrafts
are able to track their destination without deviation or with
minimal deviation from their original path.
Maneuvers that are used in the conflict resolution methods,
causes the aircraft to be diverted from the ideal and optimal
mainstream.
System efficiency metric:
measures the degree to which the aircrafts in the system are able to
follow direct and linear flight paths to their destinations.
In this paper, we define a performance criterion for each aircraft
same as follows:
Performance Metrics 21-28Semantic Cities @ AAAI 2012
Performance Metrics (cont)
here, we define a performance criterion for each aircraft
same as follows:
We try to select routes with lowest cost when we redirect the aircrafts’ main routes (i.e. the lowest deviation from the main
flight path for each aircraft)
Performance Metrics (cont) 22-28
ideal and optimal flight path for an aircraft
the amount of deviation from mainstream of aircrafts
Semantic Cities @ AAAI 2012
Performance Metrics (cont)
System Efficiency:
how much this criterion is closer to “1” indicates good performance of the system and how much this criterion is closer to “0” indicates poor performance
our proposed model for solving GCP for higher dimensions (for a great number of aircrafts that are in the congestion area) acts as a good way.
Performance Metrics (cont) 23-28
Total number of aircrafts
Performance of each aircraft
Semantic Cities @ AAAI 2012
Test Results
Table 1: Test results of applying the algorithm onto input graphs (congestion areas) with specified parameters
Test Results 24-28Semantic Cities @ AAAI 2012
Conclusion
one of the fundamental challenges in the current air traffic
management and especially in free flight method is detection
and resolution of conflicts.
in our approach, we mapped conflict detection and resolution
problem to graph coloring problem, then use imperialist
competitive algorithm to solve GCP.
Our proposed model provides an efficient and reliable
solution for solving conflicts in air traffic management
in conflict resolution process we tried to select routes with
lowest cost when the aircrafts’ redirected from main route,
therefore, we have least delay in flights and the minimum
consumption of resources (e.g. in fuel consumption).
our proposed model use multiple strategies for the resolution
of conflicts.
Conclusion 25-28Semantic Cities @ AAAI 2012
References1. Federal Aviation Regulations and Aeronautical Information Manual, 2010 edition. 2009 ASA,
Inc. Newcastle, Washington.
2. Rong, J., Geng, Sh., Valasek, J., R. Ioerger, T. 2002. Air traffic conflict negotiation and
resolution using an onboard multi agent system. Digital Avionics systems Conf. Irvine, CA.
3. Kuchar, J., and Yang, L. C. 2000. A Review of CD&R Modeling Methods. IEEE Transactions on
Intelligent Transportation Systems, Vol. 1, No. 4, December.
4. Agogino, A., and Tumer, K. 2008. Adaptive Management of Air Traffic Flow: A Multiagent
Coordination Approach. Proceedings of the Twenty-Third AAAI Conf. on Artificial
Intelligence.
5. Agogino, A., and Tumer, K. 2009. Improving air traffic management with a learning
multiagent system. IEEE Intell. Syst., vol. 24, no. 1, pp. 18–21, Jan/Feb.
6. Atashpaz-Gargari, E., Lucas, C. 2007. Imperialist Competitive Algorithm: An algorithm for
optimization inspired by imperialistic competition. IEEE Congress on Evolutionary
Computation, 4661–4667.
7. Eby, M. S. and III, W. E. K. 1999. Free flight separation assurance using distributed
algorithms. In Proceedings of aerospace conference. Aspen, Co.
8. Archibald k., Hill c., Jepsen a., Stirling c., and Frost l. 2008. A satisfying approach to aircraft
conflict resolution. IEEE transactions on systems, man, and cybernetics—part c:
applications and reviews, vol. 38, no.
References 26-28Semantic Cities @ AAAI 2012
References9. Agogino, A. K., and Tumer, K. 2009. Learning indirect actions in complex domains: Action
suggestions for air traffic control. Advances in Complex Systems, 12, 493–512.
10. Whalen, S. 2002. Graph Coloring with Genetic Algorithms.
http://mat.gsia.cmu.edu/COLOR/color.html.
11.Garey, M.R. and Johnson, D.S. 1999. Computers and intractability: a guide to the theory of
NP-completeness, W.H. Freeman and Company, New York.
12.De Werra D. 1990. Heuristics for Graph Coloring. Computing Suppl. 7, pp. 191-208.
13.Goldberg, D. E. 1980. Genetic Algorithms in Search, Optimization and Machine Learning.
Addison-Wesley, Reading, MA.
14.Valkanas, G., Natsiavas,P., Bassiliades,N. 2009. A collision detection and resolution multi
agent approach using utility functions, Fourth Balkan Conf. in Informatics.
15.Chao, W. and Jing, G. 2009. Analysis of Air Traffic Flow Control through Agent-Based
Modelling and Simulation. International Conf. on Computer Modelling and Simulation, IEEE.
16. Jensen T.R., Toft B. 1995. Graph Coloring Problems, Wiley interscience Series in Discrete
Mathematics and Optimization.
References 27-28Semantic Cities @ AAAI 2012
The End & Thanks
The End & Thanks
Thanks for your attention!