A Decomposition Heuristic for ResourceAllocationBerthe

Allocation Berthe Y. Choueiry1 and Boi Faltings2 Abstract. The allocation of resources to a set of tasks with xed start and end times is known to be NP-complete. We propose an approximation based on identifying pools of interchangeable resources and tasks using a novel decomposition heuristic. The pools are abstractions which provide simple representations for making con icts explicit. The method is an anytime algorithm that e ciently nds a sub-optimal solution which can then be improved by re ning the underlying decomposition. This is more practical than constraint relaxation techniques traditionally used to deal with such problems.