End-of-line inspection and rework are critical for achieving zero-defect manufacturing but present complex scheduling challenges due to unpredictable task durations and multiple rework cycles. This paper presents a simulation-based evaluation of a decentralized, agent-oriented scheduling framework in a Webots-based digital twin environment, modeling a quality assurance shop floor where human workers and robots work together to carry out inspection and rework tasks. Agents representing human workers, robots, and workstations dynamically pull tasks from shared task pools, with worker agents using either a First-In-First-Out heuristic or a spatial heuristic. A total of 36 manufacturing configurations were simulated, varying the number of workers, robots, and workstations. Key performance indicators such as productivity, resource utilization, and worker walking distances were analyzed. Results show that the spatial heuristic significantly reduces worker walking distances—by up to 70%—without compromising productivity, and in some cases enhances it. The study demonstrates that ergonomic and efficiency objectives can be jointly optimized and provides a scalable framework for human-centered scheduling in rework-intensive manufacturing scenarios.
Hämmerle et al. (Thu,) studied this question.