Throughput management in job shops

• Main Author: Tan, Si Pei
• Other Authors: Rajesh Piplani
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Engineering::Systems engineering; Engineering::Mechanical engineering
• Online Access: https://hdl.handle.net/10356/141011
• Institution: Nanyang Technological University

Job shop scheduling is a major concern to many manufacturing companies due to its enormous impact on throughput and cycle times of orders. Decisions are made to allocate the workload (production) between different work centers or workstations and determine the sequence in which jobs should be processed, to meet their objective. Therefore, job shop scheduling is formalized to the task of allocating finite resources over time to accomplish a given set of orders to meet one’s objective. Despite its importance, the job shop scheduling problem (JSP) remains a complex problem to solve. A semiconductor fabrication plant is one example of a highly specialized job shop. Wafers are made-to-order, and orders vary considerably in terms of the process sequencing, its requirements and setups or resources and materials needed. The scheduling of job shops is complex (Stevenson, 2011). This report aims to model jobs in a semiconductor facility to develop a heuristics-based approach and improve the scheduling of jobs to increase throughput while reducing process flow times. Using a simulation model of the semiconductor facility, the CONWIP method is evaluated to determine its effectiveness in comparison to a purely push-based system. Experiments indicate the CONWIP method is better at reducing cycle times and increasing throughput. It also regulates the cycle times and throughput better, decreasing their variability.