A multiobjective memetic algorithm for integrated process planning and scheduling problem in distributed heterogeneous manufacturing systems

With the deepening of economic globalization and the development of the manufacturing industry, distributed manufacturing patterns have become a popular topic in current production research. In the background of distributed shop scheduling, process planning problems in different factories are considered integrally with scheduling problems to utilize the heterogeneous machining resources of distributed factories. To address actual production problems more concretely, this paper investigates the multiobjective distributed integrated process planning and scheduling (MODIPPS) problem to minimize makespan, maximum machine load, and total machine load, and it establishes a mixed-integer linear programming (MILP) model. In addition, by designing a new encoding method based on the OR-nodes of the process network graph, this paper proposes a multiobjective memetic algorithm (MOMA) to solve the problem. The proposed MOMA can guarantee the feasibility of individuals by several specially designed genetic operators so that the process precedence constraints in the network graph are satisfied in the whole algorithm period. Furthermore, the algorithm introduces a simulated annealing (SA) mechanism to avoid falling into a local optimum by accepting relatively poor individuals with a certain probability. Finally, through comparison experiments on benchmarks, the proposed method shows sufficient effectiveness and superiority in solving MODIPPS problems compared with existing classic multiobjective optimization algorithms.