Nature-inspired metaheuristics for multiobjective activity crashing

Many project tasks and manufacturing processes consist of interdependent time-related activities that can be represented as networks. Deciding which of these sub-processes should receive extra resources to speed up the whole network (i.e., where activity crashing should be applied) usually involves the pursuit of multiple objectives amid a lack of a priori preference information. A common decision support approach lies in first determining efficient combinations of activity crashing measures and then pursuing an interactive exploration of this space. As it is impossible to exactly solve the underlying multiobjective combinatorial optimization problem within a reasonable computation time for real world problems, we have developed proper solution procedures based on three major (nature-inspired) metaheuristics. This paper describes these implementations, discusses their strengths, and provides results from computational experiments.