Distribution-free and Risk-averse Disassembly Line Balancing Problem
Résumé
With green and smart manufacturing emergence and development, effective conduct with End of Life (EOL) products is an extremely important and hot research topic. Disassembly line design is a fundamental and complex problem in EOL product recycling and remanufacturing, especially when considering uncertainty. In this paper, we investigate a disassembly line balancing problem (DLBP) with stochastic task processing time to minimize the cycle time. We focus on selecting alternative task processes, determining the number of open workstations and assigning operational tasks to the workstations. A joint chance-constrained programming is proposed and we adopt a distribution-free method, which is based on Markov Inequality, to approximate the chance constraints as the distribution is difficult to fully estimated. Risk-neutral model may provide solutions that perform poorly while considering uncertainty. Therefore, a risk-averse measure called Conditional Value-at-Risk (CVaR), which is rarely used in the DLBP, is considered in this paper to incorporate decision makers' perspectives. To the best of our knowledge, we are the first to consider risk-averse model in DLBP with minimizing cycle time under uncertain task processing time. Finally, computational experiments are conducted and give the impacts by testing different risk levels and confidence levels.