DPACO: Dynamic Tuning of ACO with Adaptive Strategy for QoS-Aware Web Service Composition

Abstract

With the proliferation of services available on the Web, modern systems increasingly require mechanisms ableto orchestrate multiple services in order to meet complex business requirements. Web service composition is an efficientapproach to transform these basic services into coherent composite solutions, while ensuring the scalability and flexibilityessential in distributed environments. However, Web service composition (QoS-aware WSC) based on Quality of Servicepresents a major problem. For each task in a workflow, several candidate services are available, characterized by significantheterogeneity and variability. In order to select the optimal combination of services able to satisfy QoS constraints transformsthe problem into a complex multi-objective optimization challenge that is difficult to solve in polynomial time, especially inlarge-scale and dynamic environment. In this paper, we proposed a new approach called Dynamic Parameter Ant ColonyOptimization (DPACO), an enhanced variant of the ACO algorithm for QoS-aware Web service composition. DPACOintroduces adaptive parameter control and a dynamic pheromone injection strategy, which improve the balance betweenexploration and exploitation. This adaptivity helps to avoid stagnation and local optimal, while reducing search time andensuring robustness in dynamic and heterogeneous environments. Experimental evaluations conducted on multiple datasetsdemonstrate the superiority of the proposed approach compared to classical ACO and its variants (SACO, MOACO, FACO,and EFACO). Specifically, it achieves improvements of up to +15.2% over ACO, +12.6% over SACO, +9.8% over MOACO,+7.3% over FACO, and +11.6% over EFACO in terms of solution quality. Moreover, the adaptive injection and parameteradjustment mechanisms significantly reduce the computational time required to reach optimal composite services.