A discrete differential evolution algorithm for cyclic scheduling problem in re-entrant robotic cells

Abstract

This paper addresses cyclic scheduling in robotic cells with re-entrant workstations which parts visit more than once. We present an analytical model for the problem with a given robot move sequence. Then the problem is transferred to find a robot move sequence to minimize the cycle time. An efficient discrete differential evolution (DDE) algorithm is presented to search a near-optimal robot move sequence. We encode the permutation of the robot moves as the chromosome of the individual and propose a modified mutation and crossover operations to generate the new individual. Our DDE algorithm is tested by a numerical instance.