Efficient optimization of robust project scheduling for industry 4.0: A hybrid approach based on machine learning and meta-heuristic algorithms

Abstract

This research contributes significantly to the domain of Industry 4.0 by offering a nuanced approach to the multi-objective optimization of the resource-constrained project scheduling problem (RCPSP) under uncertainty. Focused on the context of smart product platforming, this study introduces a novel methodology that not only considers traditional factors like time and cost but also incorporates quality and risk aspects, crucial for personalized product fulfillment. In this regard, a comprehensive four-objective mathematical model is proposed to minimize project completion time, total project costs, and project risks while simultaneously enhancing overall project quality. Real-world uncertainty is acknowledged through the incorporation of uncertain parameters for the time, risk, and quality associated with each project activity. To address this uncertainty, a robust optimization method is applied based on Bertsimas and Sim's approach. Moreover, to optimize the proposed model, the Hybrid Red Deer and Genetic Algorithm (HRDGA) is proposed, which is leveraging a machine learning approach for clustering solutions. The numerical results demonstrate that increasing the project budget by 30% leads to an upward trend in total project costs and a reduction in the minimum acceptable quality by 10%–30% results in a decreasing trend in the total project cost. This research emphasizes the adoption of Industry 4.0 enabling technology within the project scheduling platform, particularly highlighting its significance for personalized product fulfillment.