Constraint Handling Methods for Resource-Constrained Robotic Disassembly Line Balancing Problem

Abstract

Abstract In this paper, we analysed a resource-constrained robotic disassembly line balancing problem (RCDLBP), which was significantly different from traditional DLBP for the robots will consume a certain amount units of resources when performing disassembly tasks. These resources could be some scarce disassembly tools or the oil and electricity to support the normal work of robots. A mathematical model of RCDLBP which considering an additional resource constraint and simultaneously minimizes the cycle time and the number of robots used was proposed in this research. Based on the mathematical model and three constraint handling methods which integrated with multi-objective evolutionary algorithm based on decomposition (MOEA/D), we performed experiments on 16 test cases. The initial population feasible ratio of all test instances ranges from 99.99% to 0.34%. Results show that as the initial population feasible ratio gradually decreases, the search capabilities of different constraint handling methods changed dramatically.