hxj0316/ant_colony_algorithms_for_multi_agents

this is a multi-agent algorithm in task planning level based on ant colony algorithms.

Ant Colony Algorithms for Multi-agent talk planning

build

catkin_make

execute

roscore
rosrun acolanner aco_planner_talker
rosrun acolanner aco_planner_listener
rosrun aco_planner aco_planner_node

explanation

(1) aco_planner_talker
publish all the task points that multiple agents need to arrive.

(2) aco_planner_listener
subscribe the result that ant colony algorithms compute.

(3) aco_planner_node
core code to compute the result based on ant colony algorithms.

parameters

the map:
(now is a blank map)

m_srcImg = cvLoadImage("src/aco_planner/src/m_planImg.jpg",-1);         //把原始图形读进m_srcImg

the task nodes:

std::string full_path = ros::package::getPath("aco_planner") + "/src/Nodes.txt";

results

the task points that each agent need to go:

the planning path for each agent need to go, and different color indicates different agent:

while other process files can be found in src/aco_planner/Debug folder.

Thanks

Friend fshi2006.