Heuristic algorithms to minimize total weighted tardiness with stochastic rework and reprocessing times

• Studied a scheduling problem with stochastic rework and reprocessing times in electronic manufacturing systems.
• Developed a mathematical model to capture different system parameters including in- and off-line reworks.
• Developed two greedy heuristics based on Total Estimated Processing Time (TEPT).
• Showed that the proposed algorithms outperform compared to different dispatch rules in terms of solution quality and computation time.

This paper introduces a mathematical model for the scheduling problem with stochastic rework and reprocessing time, which is typical for electronics manufacturing services (EMS) providers. Since rework and reprocessing of jobs take more time resulting in missing the due dates, this research aims to determine the optimal job sequence on the machines. A heuristic methodology is developed that takes into consideration the Total Estimated Processing Time (TEPT), a linear combination of processing, rework, and reprocessing times. Jobs with different configurations of processing, rework, and reprocessing times representing a High-Mix–Low-Volume (HMLV) setup are tested on a single-machine job shop system to understand the effectiveness of the shortest TEPT algorithm (STEPT). Improvements to the STEPT algorithm are made and the modified STEPT algorithm (MSTEPT) is tested using a single-machine job shop setup with a larger number of jobs. Experimental results indicate that the proposed algorithm outperforms different commonly used dispatch rules in terms of solution quality and computation time. Then, experiments conducted on a multi-machine job shop setup with larger number of jobs also indicate the superior performance of the MSTEPT algorithm.