Manufacturing Line Optimization Using Stochastic Models Under Uncertain Conditions

Abstract

The uncertainty around manufacturing organizations is growing with failures in the machines, changing customer orders, unpredictable processing times, maintenance downtimes and limited resources. All these uncertainties have a considerable impact on production efficiency, throughput, system reliability and operational decision making. Most traditional deterministic optimization methods are not well suited to model the dynamics of modern manufacturing systems and therefore they are not useful to solve production problems in practice. In this study, manufacturing line optimization under uncertain conditions using stochastic models is investigated and an integrated framework is proposed which incorporates the theories of queueing theory, Markov chain analysis and simulation based evaluation. The research explores how uncertainty affects manufacturing performance, and investigates optimization methods that can enhance the operational results. A stochastic production model is created to capture all the variations in machine behaviour, processing times and production flows. Markov chain modelling is used for machine reliability and state transition analysis and queueing analysis is used for identifying the bottlenecks and wait-time characteristics. Multiple uncertainty scenarios for simulation experiments are performed to study throughput, utilization, queue length, and system availability. The results show that stochastic optimization models are more realistic descriptions of manufacturing operations and are more effective at enhancing production performance than the traditional deterministic ones. The results show the reduction in waiting time and queue size, increased throughput, machine usage, and operational resilience. In addition, scenario analysis has been used to show that uncertainty management positively impacts flexibility and sustainability in manufacturing. Based on the conclusion of the study, stochastic modelling is proved as a useful decision support tool for optimizing manufacturing systems in uncertain situations, and also the practical advantages of the stochastic modelling in production planning, resource allocation, maintenance management and strategic decision making in the modern industrial environment are highlighted.