An Optimization of multi-level multi-objective cloud production systems with meta-heuristic algorithms

The inability of small companies to compete with large production systems has led to the sharing resources in cloud production systems among smaller companies to compensate for their production deficiencies. This study proposes a multi-objective, multi-level cloud production system by minimizing the maximum completion time of activities, the costs of the entire cloud production system, and the maximum risk of information disclosure. We use a support vector machine (SVM) to train the input data, including activities, micro activities, and services of production units. The correlation between the trained and predicted data from the machine learning model equals 0.9977, indicating this method’s high accuracy and efficiency. The Lp-Metric, multi-objective grey wolf optimizer (MOGWO), and non-dominated sorting genetic algorithm-II (NSGA-II) algorithms were used to solve the problem using trained input data. The Lp-Metric results show that reducing the completion time of all activities and the maximum risk of information disclosure increases the costs of the entire cloud production system. We also examine the efficiency of the solution methods and demonstrate that MOGWO is more efficient in solving the cloud production system problem. • Allow small companies to compete with large production systems by sharing resources in cloud production systems. • Propose an optimization of multi-level, multi-objective cloud production systems with meta-heuristic algorithms. • Minimize the maximum time to complete activities, the total costs of the cloud production system, and the maximum risk of information disclosure. • Use support vector machines to train the input data, including activities, micro activities, and services of production units. • Use Lp-metric, multi-objective grey wolf optimizer (MOGWO), and non-dominated sorting genetic algorithm-II (NSGA-II) to optimize objective functions.