A Distributed Approach To Solving Constrained Multiagent Task Scheduling Problems

The constrained multiagent task scheduling problem is a problem class that describes complex scheduling problems that integrate stochastic task allocation with constraint satis-faction for agents that concurrently execute multiple tasks. The combination of elements from constraint satisfaction, task allocation and resource scheduling produces a challeng-ing class of problems. Unfortunately, existing work in multi-agent scheduling limits the solution space and therefore task allocations do not completely encompass the complexity and interaction of the full multiagent task scheduling problem. We propose a problem description and distributed approach designed to directly solve such problems allowing optimal and approximated solution comparison. Experimentation us-ing a simulated alarm handling domain provides empirical ev-idence of the successful adaptation of our approach to solving constrained multiagent task scheduling problems with scala-bility towards large problem sizes