A Business Process Management System based on a General Optimium Criterion

Abstract

Business Process Management Systems (BPMS) provide a broad range of facilities to manage operational business processes. These systems should provide support for the complete Business Process Management (BPM) life-cycle [16]: (re)design, configuration, execution, control, and diagnosis of processes. BPMS can be seen as successors of Workflow Management (WFM) systems. However, already in the seventies people were working on office automation systems which are comparable with today’s WFM systems. Recently, WFM vendors started to position their systems as BPMS. Our paper’s goal is a proposal for a Tasks-to-Workstations Assignment Algorithm (TWAA) for assembly lines which is a special implementation of a stochastic descent technique, in the context of BPMS, especially at the control level. Both cases, single and mixed-model, are treated. For a family of product models having the same generic structure, the mixed-model assignment problem can be formulated through an equivalent single-model problem. A general optimum criterion is considered. As the assembly line balancing, this kind of optimisation problem leads to a graph partitioning problem meeting precedence and feasibility constraints. The proposed definition for the "neighbourhood" function involves an efficient way for treating the partition and precedence constraints. Moreover, the Stochastic Descent Technique (SDT) allows an implicit treatment of the feasibility constraint. The proposed algorithm converges with probability 1 to an optimal solution.