Job scheduling and collision resolution of multi-bridge processing systems

Abstract

A multi-bridge processing system (MBPS) has multiple bridge machines working at their serially arranged and partially overlapping workspaces. Its job scheduling can be abstracted as a colored travelling salesman problem (CTSP). CTSP is a new type of multiple traveling salesman problems in which each salesman must visit his exclusive cities and may visit some shared cities. This work presents a two-stage method for scheduling multi-bridge jobs first and resolve inter-bridge collision next. First, it proposes a population-based incremental learning (PBIL) algorithm to solve a serial-CTSP and the job scheduling problem of MBPS. PBIL processes two types of possibility vectors for city assignment/selection and a local search operation for improving its search ability. Second, this work designs a mechanism of collision resolution to remove potential inter-bridge collision contained in the obtained schedule. Finally, this work applies the proposed method to a triple-bridge waterjet cutting process to show its validity.