Industrial edge cloud deployment algorithm for industrial internet of things

If the industrial edge cloud deployment covering the production line is unreasonable, it is easy to cause problems such as the decline of real-time operation and maintenance service quality and the increase of enterprise costs under the dynamic adjustment of production lines in different regions and the limited cloud resources. By using the idea of constrained multi-objective programming and constrained minimum subset partition, the industrial edge cloud deployment problem was discussed, and a heuristic genetic algorithm was proposed. Based on the characteristics of the problem, the binary coding was adopted to reduce the difficulty of algorithm implementation. The multi-round random non-repetitive solution strategy was used to select the feasible solutions as the initial population, so as to improve the search speed and search probability. According to the mixed selection method, the better individuals and the worse individuals were selected purposefully to maintain the diversity of the population. The method of multi-round, multi-dimension and multi-point crossover was adopted to realize the deep crossover of better and better individuals, better and worse individuals, worse and worse individuals, to maintain population diversity and to explore new areas. The preferred individual preference single-point mutation strategy was adopted to carry out local deep excavation in the region where the preferred new individuals generated by cross-operation were located, and the direction of deep excavation was adjusted continuously in the process of deep excavation, so as to maintain the diversity of population and enhance the global search ability. The validity, convergence and global search ability of the algorithm were verified from expectation load deviation rate, expectation service delay deviation rate, algorithm convergence rate and solutions error rate. © 2022, Editorial Department of CIMS. All right reserved.