Work

Bi-Objective Parallel Batch Scheduling with Two-Level Incompatibility and Asymmetric Sequence-Dependent Setup Times: A Sustainable Textile Dyeing Application

Year: 2026
Type: article
Abstract: Cleaning dye vessels between colour changes consumes water, chemicals, and energy. We study its trade-off with on-time delivery in a bi-objective parallel batch scheduling problem with two simultaneous compatibility restrictions, colour family and customer, and asymmetric sequence-dependent setup times; total setup time serves as the proxy for cleaning. We give a mixed-integer formulation, show that each criterion is strongly NP-hard, and propose a two-stage metaheuristic that forms batches and then sequences them by a tabu search coupled to the ε-constraint bound, generating approximation fronts (sets of mutually non-dominated schedules). On 120 instances, each solved five times under equal computing time, the method attains the highest hypervolume against adaptive large neighbourhood search and the non-dominated sorting genetic algorithm II (NSGA-II) on every instance, and on 115 when normalised by a current-practice dispatching rule. Its hypervolume margin over NSGA-II grows from 0.10 at 30 to 0.24 at 100 jobs, and it is larger still without the annealing step of its batching stage. Against the dispatching rule, its fronts offer on average 42% less tardiness without more setup time, or 86% less setup time without more tardiness. On the 11 small instances (up to 10 jobs) with a known optimum, its tardiness-minimal schedule is optimal; other points are not certified.
Source: Mathematics
Language: English
Open Access status: gold
Estimated APC paid: $3,137
OpenAlex ID: W7236819600
FWCI:
Cites: 75
Cited by:
Related to: 10