A two-phase approach for benefit-driven and correlation-aware service composition allocation in cloud manufacturing

Abstract

Manufacturing service composition (MSC) is a fundamental component of cloud manufacturing that involves multiple stakeholders and associated services. Stakeholders, viewed as autonomous entities, prioritize both temporary gains and long-term benefits, while service interactions significantly influence the feasibility and quality of MSCs. Effective MSC allocation demands a strategic approach that harmonizes stakeholders’ interests and accounts for service correlations to achieve optimal outcomes. Although previous studies have explored stakeholders’ long- and short-term benefits and service correlations independently, a comprehensive framework that integrates these aspects, aligns diverse interests, and thoroughly examines service correlation impacts remains a significant challenge. To bridge this gap, this study proposes a two-phase approach for benefit-driven and correlation-aware MSC allocation. This approach integrates the long- and short-term benefits of platform operators, service requesters, and resource suppliers, as well as the effects of two types of service correlations, composability-focused and quality-focused, on MSCs. The initial phase constructs a bi-objective optimization model with maximizing operational benefits to filter and recommend MSCs. Composability-focused service correlations are incorporated in the model to ensure MSC feasibility. The second phase refines the optimal MSC selection by matching supply and demand, leveraging quality-focused service correlations to modify MSC quality. A hybrid algorithm, combining an improved fast nondominated sorting genetic algorithm (PANSGA-II) with an enhanced technique for order preference by similarity to the ideal solution (TOPSIS-PR), is developed to solve the two-phase problem. The case study and numerical experiments are conducted to validate the applicability and effectiveness of the proposed approach and algorithm.