Model maturity-based model service composition in cloud environments
Ying Liu,Lin Zhang,Yongkui Liu,Yuanjun Laili,Weicun Zhang
DOI: https://doi.org/10.1016/j.simpat.2021.102389
IF: 4.199
2021-12-01
Simulation Modelling Practice and Theory
Abstract:With the development of cloud computing (CC), service-oriented architecture (SOA), and container technology, modeling and simulation (M&S) resources, such as simulation software and different sorts of models, can be shared and reused in a cloud environment. Modeling and Simulation as a Service (MSaaS), as a new paradigm, supports sharing simulation models or modeling tools and has enabled a wide range of model reuse. However, reusing or combining some immature models may result in inefficient M&S activities or even false simulation results. To make sure the appropriate reuse and composition of simulation models in cloud environments, which is also termed as model service composition for simulation (MSCS), this paper incorporates model maturity with service cooperation as a metric to evaluate the quality of model composition in cloud. Then, as a multi-objective optimization problem with multiple constraints, the MSCS problem and its process are described in detail. To solve the MSCS problem, a novel evolutionary algorithm named CA-AO-NSGAII is proposed. In the algorithm, adaptive crossover and mutation operators, as well as probabilistic initialization are developed. Furthermore, a half-local search algorithm in an elitist mechanism is designed for efficient decision-making. To validate the performance of CA-AO-NSGAII, experiments with respect to four different cases are conducted. Results show that the proposed method for addressing MSCS issue is effective and feasible.
computer science, interdisciplinary applications, software engineering