Adaptability of multirecombinated evolutionary algorithms to changing common due dates

Abstract

In the restricted single-machine common due date problem, the goal is to find a schedule for the n jobs which jointly minimizes the sum of earliness and tardiness penalties. This problem, even in its simplest formulation, is an NP-Hard optimization problem. New trends to enhance evolutionary algorithms introduced multiple-crossovers-on-multiple-parents (MCMP) a multirecombinative approach allowing multiple crossovers on the selected pool of (more than two) parents. MCMP-V is a novel MCMP variant, which directly applies multirecombination to the Lee and Kim approach using uniform scanning crossover.