Two-stage assembly scheduling problem for processing products with dynamic component-sizes and a setup time. (February 2017)

• Record Type: Journal Article
• Authors: Jung, Sunwoong; Woo, Young-Bin; Kim, Byung Soo
• Is Part Of: Computers & industrial engineering. Volume 104(2017)
• Page Start: 98
• Page End: 113
• Publication Date: 2017-02
• Subjects: Scheduling -- Two-stage assembly flow shop system -- Setup time -- Mixed integer linear programming (MILP) -- Genetic algorithm; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2016.12.030
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: We study a scheduling problem of two-stage assembly flow-shop system. The system processes products with dynamic component-sizes and a setup time. We simultaneously determine production scheduling and assembly scheduling. We derive a novel mixed integer programming model for the problem. We propose a good hybrid GA with a local search heuristic. Abstract: In this paper, two-stage assembly flow shop scheduling problem (TSAFSP) to assemble products having dynamic component-sizes is considered. In the machining stage, a single machining machine produces various types of components to assemble the products. During the machining process, a setup time is required whenever the machining machine starts to process a new component or processes a different component. When the required components are available for the associated product from the machining stage, a single assembly machine can assemble these components into the product in the assembly stage. To solve the problem, a novel mixed integer linear programming model is derived. Three genetic algorithms (GAs) with different chromosome representations are proposed due to the intractability of the optimal solution for large-sized problems. One GA has a chromosome to represent a complete solution. Two hybrid genetic algorithms (HGAs) have a simple chromosome to represent a partial solution, and the rest of the solution is provided by an effective local search heuristic given the partial solution. The performance of the GAs is compared by using randomly generated examples.