The HMT Laboratory’s multiple robot/agent task planning research goal is to investigate integrating coalition formation and planning in order to develop a decision support system for multiple heterogeneous agent systems. Coalition formation partitions a set of agents into teams and assigns them to tasks. Planning determines a sequence of actions to take to accomplish a task based on a task specification and a set of agents. Current research treats these as two independent problems; however, these two problems are tightly coupled in real-world systems. The groupings of agents produced by coalition formation determine which agents are available for the planning system.

While valid solutions are found by solving the two problems separately, considering the two problems simultaneously will provide a better solution. Instead of forming coalitions and then planning for each task individually, all agents will be considered when planning for all tasks to allow task and agent interactions to be considered in the developed plan. Considering all agents when planning increases the difficulty of the problem by increasing the space of possible plans, thus new algorithms will be developed that draw from existing coalition formation and planning algorithms.