MDE based Memetic Algorithm using OpenMP and its application in engineering project scheduling problems with dynamic due dates

A multi-population DE based Memetic Algorithm using OpenMP is proposed, and applied to a kind of engineering project problems which can be modeled by single machine total weighted tardiness problems with dynamic due dates. The multi-population scheme and a best-survival individual synchronous updating strategy are introduced to simultaneously refresh the sub-populations in the same thread. The searching ability of the algorithm is improved by the differential evolutionary Memetic algorithm which is formulated by incorporating a local search method for the best individual in each sub-population into DE. The effectiveness of the algorithm is proved by a series of engineering project scheduling based simulation experiments.