Multi-Objective Batching-Dispatching Framework for Parallel Machine Scheduling with Job Families

Under review · Computers and Industrial Engineering (C&IE), submitted 2025

Authors: Karen Seojin Kim†, Minju Jang†, Sang-Wook Lee, Jimin Park, and Hyun-Jung Kim  († Co-first authors) Status: Under review in Computers and Industrial Engineering (C&IE), submitted in 2025


TL;DR

When jobs belong to families that share a common setup, processing them in batches can save a lot of machine time — but only if batching and machine assignment are decided together. We propose a framework that couples batching and dispatching for parallel machine scheduling and optimizes several objectives at once.

The problem

In parallel machine scheduling with job families, switching a machine from one family to another incurs a setup. Grouping same-family jobs into a batch avoids repeated setups, but larger batches can delay individual jobs. The scheduler must therefore trade off setup efficiency, throughput, and lateness — and do so across multiple machines simultaneously.

The approach

  • Batching + dispatching, jointly — instead of forming batches first and assigning them later, the framework decides how to batch and where/when to dispatch in a coordinated way.
  • Multi-objective — it balances competing goals (e.g. makespan, tardiness, and setup count) rather than optimizing a single metric.

Why it matters

Family setups dominate cost in many real lines (semiconductor, electronics, process industries). A coordinated batching–dispatching policy gives planners a practical way to cut setup time without sacrificing due-date performance.


My Role

I contributed to this work as a co-author.

✍️ Feel free to refine this summary with the exact abstract, objectives, solution method, and results, and to describe the specific part you contributed.