Game-theoretic planning

One common problem in multi-agent planning is when strategic agents need to execute a plan in a shared environment and the execution of such a plan may interfere with the plan of others. In this case, agents are competitors but are also willing to cooperate by accommodating their plans to the others' and thus ensure a minimal profit. Agents have alternative plans for their purposes, each associated to a different utility value, but they can also schedule a plan at a different time to avoid collisions with the other agents’ plans, which in turn may also entail a loss of utility.

In this research line, we are developing a collaborative game-theoretic approach that finds a global stable solution in Nash Equilibria for this problem, where collaboration is viewed as each agent taking the best response to the strategic choices of the other agents.