Accommodating Plan Revisions with Multiple Agents for Local Search in Road Networks

In planning operations involving searching all nodes in a networks, one is interested in determining an optimal search. Optimal searches aim to use all available agents to find the quickest allocations and paths to search all nodes. However, problems arise when an agent is dropped during the search process; for example in a search and rescue mission an Unmanned Aerial Vehicle (UAV) might be tasked to stay at a location after finding survivors while the remaining agents search the rest of the map. When such situations occur, plan revision must be used to re-plan the search on the remaining unsearched nodes with the remaining agents. This work develops various plan revision strategies with application to two real road networks. Results are considered and evaluated for varying numbers of agents.