Integrated process planning and scheduling with setup time consideration by ant colony optimization

Abstract

In recent years, lots of research effort was spent on the integration of process planning and job-shop scheduling. Various integrated process planning and scheduling (IPPS) models and solution approaches have been proposed. The previous and existing research approaches are able to demonstrate the feasibility of implementing IPPS. However, most of them assumed that setup time is negligible or only part of the processing time. For machined parts, the setup for each operation includes workpiece loading and unloading, tool change, etc. For setup that depends only on the operation to be processed (sequence-independent), it is applicable to adopt the assumption of not considering setup in IPPS. For setup that depends on both the operation to be processed and the immediately preceding operation (sequence-dependent), it is an oversimplification to adopt such assumption. In such cases, the setup time varies with the sequence of the operations. The process plans and schedules constructed under such assumption are not realistic or not even feasible. In actual practice, therefore, the setup time should be separated from the process time in performing the IPPS functions. In this thesis, a new approach is proposed for IPPS problems with setup time consideration for machined parts. Inseparable and sequence-dependent setup requirements are added into the IPPS problems. The setup times are separated from the process times and they vary with the sequence of the operations.

IPPS is regarded as NP-hard problem. With the separated consideration of setup times, it becomes even more complicated. An Ant Colony Optimization (ACO) approach is proposed to handle this complicated problem. The system is constructed under a multi-agent system (MAS). AND/OR graph is used to record the set of feasible production procedures and sequences. The ACO algorithm computes results by an autocatalytic process with the objective to minimize the makespan. Software agents called “artificial ants” traverse through the feasible routes in the graph and finally construct a schedule. A setup time parameter is added into the algorithm to influence the ants to select the process with less setup time. The approach is able to construct a feasible solution with less setup time.

Experimental studies have been performed to evaluate the performance of MAS-ACO approach in solving IPPS problems with separated consideration of setup times. The experimental results show that the MAS-ACO approach can effectively handle the problem.