Task optimization and scheduling of distributed cyber-physical system based on improved ant colony algorithm

Abstract Cyber–physical system (CPS) is the product of technological development to a certain stage, and also is the future trends in information technology. High-performance computing ability is the guarantee of CPS’s real-time and accuracy applications, and the emergence of distributed technology provides the implementation possibility of high-performance CPS. Task scheduling is a typical combination optimization problem and the task allocation problem on multi-processor distributed systems refers to how to use system resources most efficiently in a distributed computing environment to complete a limited set of tasks. Based on the behavior of ants searching for food in nature, ant colony algorithm is a kind of positive feedback algorithm with good robustness and easy parallel implementation and has certain advantages for dealing with constraint satisfaction. In order to introduce an adaptive mechanism and mutation strategy, shorten the calculation time of ant colony algorithm, speed up CPS algorithm convergences, and improve distributed CPS prediction accuracy, this paper analyzed the research status and significance of ant colony algorithm, expounded the development background, current situation, and future challenges of task optimization and scheduling of distributed CPS, elaborated the principles and methods of ant colony optimization algorithm model and mathematical description of CPS task scheduling, proposed a task management model of distributed CPS based on improved ant colony algorithm, explored the task optimization scheduling of distributed CPS based on improved ant colony algorithm, and finally conducted an numerical simulation to test the effect the proposed algorithm and model. The simulation results show that the proposed algorithm model enhances the local search ability and improves the quality of the task scheduling problem, and has good effectiveness, stability and adaptability. The study results of this paper provide a reference for the further research on the optimization and scheduling of distributed CPS tasks.