Development of Agent-Based Heuristic Optimisation System for Complex OEM Flow-Shop under Customer-Imposed Production Disruptions

Abstract

Original Equipment Manufacturers (OEMs) flow-shop systems are exposed to production disruptions caused by automotive assembly customers. When the customer assembly line experiences uncertainties, demand requirements change. Therefore, the problem extends to affect the OEMs flow-shop production planning and scheduling. The continuous customer changing demand requirements in terms of quantity, sequence and time of delivery of orders create disruptions. The combination of these types of disruptions on OEM flow-shop makes the problem complex to solve, hereby requiring a more robust approach, especially in an environment where customers’ demand satisfaction is prioritised, despite disruptions. In this research, a new and innovative disruption-resolution framework is proposed to tackle customer-imposed disruptions in OEMs flow-shop. The framework integrates the dynamics of agent-based simulation, inventory control, and adaptive heuristic algorithm. The heuristic algorithm is proposed to specifically help OEM flow-shop adapt and accommodate disruptions through an innovative inventory ‘borrow and replenishment’ strategy for production support when disruptions occur. The autonomous capability of agent-based simulation was adopted for simulating the actions and interactions of flow-shop resources (agents) for better system assessment of the system. As production resource such as operators and machine play a vital role in performance improvement, agent-based method is adopted to simulate system resources interaction, to improve productivity further Based on combination of disruptions occurrences, the research conducted different scenario experiments using real-life data to verify and validate the proposed approach. The results of the proposed approach, in terms of selected Key Performance Indicators (KPIs) showed an improved performance of resources by 6.89%, which led to increased number of orders satisfaction, reduced number of late/unsatisfied orders by 8.28% improvement, compared to both a sequential replenishment approach and the current flow-shop operation (“As-Is”). This showed the effectiveness of the proposed framework approach to solving the OEMs flow-shop customer-imposed disruptions problem of this nature.

Date of Award	Jun 2020
Original language	English
Awarding Institution	Coventry University
Supervisor	Ammar Al Bazi (Supervisor)