RoboCup Rescue Simulation

I have been the member of Impossibles team ( RoboCup Rescue Simulation ) since It was Initiate on March 2004.

2^nd Place in AUT-Open. Tehran, Iran, May 2005.
World championship and 1^st Place in International RoboCup Rescue Simulation League, Osaka Japan, July 2005.
1^st Place in AI-Game2005.Esfehan, Iran, Dec 2005.

Assign Specific Region for each Police Force Agent.
Match critical Blockade to agents so never more than one police force on a blockade.
Critical blockades are the blockades that make tow region of city unreachable to each other. It decide using graph model of city.
If the whole city is reachable from anywhere each police force will clear blockade in his/her region in order of its priorities. It is possible that a blockade is never cleared because of its low priority.
During their work they search for civilians using search models and reporting the condition of civilians that were found before.

The Communication system is completely independent from the action of agents and has been implemented in lower level of abstraction.
The Communication packets just contain row data and no processed date is sent or received.
Each side of communication has approximate world model of other side in order to send appropriate information for that side.
Each center sends n/2 packet to other centers and n packet to its agents. (2n packet)

Searching is not assigned to special agent so each agent will search in her idle times.
If we model the buildings Graph that buildings are vertexes and tow building are connected if and only if agents can see other building from one of them. So a dominating set of this graph is enough to search in order to search all the maps. As minimum set problem is an NP-Completeness problem I must solve it offline or approximately using:

Offline: dominating set of each find and store in a file using an offline program. (set of buildings) so it is enough to check these building during game.
Online: for random map it calculated during the running time so when an agent checks a building so its neighbors (in graph) were removed from graph.