Abstract
In the disassembly line balancing problem, the disassembly time of task is usually uncertain due to the influence of various factors. Interval number theory is very suitable to solve this problem. In this paper, a new interval mathematical model is proposed and the objectives are to minimize the cycle time and the total energy consumption of robots. To solve this problem, an evolutionary algorithm named γ based-NSGA-II for the interval multi-objective optimization is proposed. This algorithm directly solve the original interval multi-objective optimization problem by using interval Pareto dominance and interval crowding distance, rather than transforming the problem into a determined parameter optimization problem, which can retain the uncertain information, making the solution more reliable. And the local search operator is proposed to strength the local search ability of the algorithm. Experiment is executed in the three scale problems. By comparing the value of HV-U and HV-D, the influence of γ on the convergence, distribution and uncertainty of the algorithm is analyzed, and the optimal value of γ for this problem is found. On this basis, the performance of the proposed γ based-NSGA-II is compared with NSGA-II and MOEA / D by the value of IGD. The results show that the proposed algorithm has good performance in the small and medium scale problems.