Energy-efficient open-shop scheduling with multiple automated guided vehicles and deteriorating jobs

Abstract

There is growing interest in energy-efficient production scheduling research because of the increasing energy shortage. However, most existing studies along this line of research have not considered the energy consumed by automated guided vehicles (AGVs) used in modern smart factories for production scheduling purposes. In this paper, we study an energy-efficient open-shop scheduling problem with multiple AGVs and deteriorating jobs. A multi-objective model with four objectives is formulated, aiming to simultaneously minimise the maximum ending time of all AGVs, the total idle time of machines and AGVs, the total tardiness of jobs, and the total energy consumption of machines and AGVs. An improved population-based multi-objective differential evolution (IMODE) algorithm is developed to solve the problem. The IMODE makes use of a problem feature-based heuristic and a mean entropy method to enhance the diversity of its initial population. A novel grey entropy parallel analysis-based fitness evaluation mechanism with reference points is adopted to evaluate the candidate solutions. To improve the local search ability of IMODE, a multi-level local search strategy is used. In the experimental study, Taguchi analysis is employed to obtain the best parameter combination. The effects of the main components of IMODE are validated via comprehensive comparison experiments. Extensive experimental results show that the IMODE is preferable to other well-known multi-objective algorithms at solving the problem being considered.