Dynamic vehicle routing using Ant Based Control Ronald
Dynamic vehicle routing using Ant Based Control Ronald Kroon Leon Rothkrantz Delft University of Technology October 2, 2002 Delft Mediamatics / Knowledge based systems
Contents l Introduction l Theory l Ant Based Control l Simulation environment and Routing system l Experiment and results l Conclusions and recommendations Mediamatics / Knowledge based systems 2
Introduction (1) Dynamic vehicle routing using Ant Based Control: Routing cars through a city l Using dynamic data l Using an Ant Based Control algorithm l Mediamatics / Knowledge based systems 3
Introduction (2) Goals: Design and implement a prototype of dynamic Routing system using Ant Based Control l Design and implement a simulation environment for traffic l Test Routing system l Mediamatics / Knowledge based systems 4
Introduction (3) Possible applications: Navigate a driver through a city l Find the closest parking lot l Divert from congestions l Mediamatics / Knowledge based systems 5
Mediamatics / Knowledge based systems 6
Mediamatics / Knowledge based systems 7
Schematic overview of the PITA components Mediamatics / Knowledge based systems 8
3 D Model of dynamic traffic data Mediamatics / Knowledge based systems 9
Theory (1) Natural ants find the shortest route Mediamatics / Knowledge based systems 10
Theory (2) Choosing randomly Mediamatics / Knowledge based systems 11
Theory (3) Laying pheromone Mediamatics / Knowledge based systems 12
Theory (4) Biased choosing Mediamatics / Knowledge based systems 13
Theory (5) 3 reasons for choosing the shortest path: Earlier pheromone (trail completed earlier) l More pheromone (higher ant density) l Younger pheromone (less diffusion) l Mediamatics / Knowledge based systems 14
Ant Based Control (1) Application of ant behaviour in network management Mobile agents l Probability tables l Different pheromone for every destination l Mediamatics / Knowledge based systems 15
Ant Based Control (2) Probability table 1 2 3 (Node 2) 5 7 Mediamatics / Knowledge based systems 1 3 5 1 0. 90 0. 02 0. 08 3 0. 03 0. 90 0. 07 4 0. 44 0. 19 0. 37 5 0. 08 0. 05 0. 87 … … Destination 6 4 Next 16
Ant Based Control (3) Forward agents Generated regularly from every node with random destination l Choose route according to a probability l Probability represents strength of pheromone trail l Collect travel times and delays l Mediamatics / Knowledge based systems 17
Ant Based Control (4) Backward agents Move back from destination to source l Use reverse path of forward agent l Update the probabilities for going to this destination l Mediamatics / Knowledge based systems 18
Ant Based Control (5) Updating probabilities l Probability for choosing the node the forward agent chose is incremented Depends on: • Sum of collected travel times • Delay on this path Update formula: Δp = A / t + B l Probabilities for choosing other nodes are slightly decremented Mediamatics / Knowledge based systems 19
Simulation environment and Routing system (1) Architecture Simulation GPS-satellite Vehicle Routing system Mediamatics / Knowledge based systems 20
Simulation environment and Routing system (2) Communication flow GPS-satellite • Position determination Vehicle • Routing • Dynamic data Routing system Mediamatics / Knowledge based systems 21
Routing system (1) Routing system Dynamic data Timetable updating system Memory Mediamatics / Knowledge based systems 22 Route finding system Routing
Routing system (2) Timetable 1 2 3 1 2 4 5 … 6 4 5 7 Mediamatics / Knowledge based systems 23 1 2 4 - 12 15 5 … - … 11 - - 18 … 14 - - 13 … - 18 14 - … … …
Routing system (3) Update information t 1 1 3 2 t 2 20 6 4 5 7 Mediamatics / Knowledge based systems 24
Mediamatics / Knowledge based systems 25
Simulation environment (1) Map of Beverwijk Mediamatics / Knowledge based systems 26
Simulation environment (2) Map representation for simulation Mediamatics / Knowledge based systems 27
Simulation environment (3) Simulation with driving vehicles Mediamatics / Knowledge based systems 28
Simulation environment (4) Features l l l Traffic lights l Roundabouts l One-way traffic l Number of lanes l High / low priority roads Mediamatics / Knowledge based systems Precedence rules Speed variation per road Traffic distribution Road disabling 29
Experiment Mediamatics / Knowledge based systems 30
Results In this test case (no realistic environment): 32 % profit for all vehicles, when some of them are guided by the Routing system l 19 % extra profit for vehicles using the Routing system l Mediamatics / Knowledge based systems 31
Conclusions l Successful creation of Routing system and simulation environment l Test results: – Routing system is effective: Smart vehicles take shorter routes l Other vehicles also benefit from better distribution of traffic l – Routing system adapts to new situations: l 15 sec – 2 min Mediamatics / Knowledge based systems 32
Recommendations l Let vehicle speed depend on saturation of the road l Update probabilities using earlier found routes compared to new route l Use the same pheromone for all parkings near a city center Mediamatics / Knowledge based systems 33
Start demo Demo Mediamatics / Knowledge based systems 34
- Slides: 34