Using a Compositional Function Hybridization of GCDPSO and GA to Solve Task Allocation Problems in Smart Farms

Compared with cloud computing, edge computing has the advantages of low latency, reduced network pressure and network risk. As a result of its development, smart farms have also started to shift from cloud services-oriented to edge computing-oriented, which is to put the tasks of smart farms on edge servers. For some reasons, such as energy limitation, the number of tasks that the edge server can handle is limited in a period of time, so some tasks still need to be processed by the cloud server. When these tasks need to be processed not only by edge servers but also by cloud servers, they need to be divided into two parts, one part which requires as much cloud processing cost as possible is processed on the edge server, and the other part which requires as little cloud processing cost as possible is processed on the cloud server. In our model, task is divisible. To solve this task assignment problem, this study considered a functional composition integration approach, Hybridization of GCDPSO and GA (HGAGCDPSO), where GCDPSO is Guaran-teed Convergence Discrete Particle Swarm Optimization and GA is Genetic Algorithm. In this paper we compares HGAGCDPSO performance with GCDPSO, Greedy Algorithm and Dynamic Programming. Results revealed that the overall performance of HGAGCDPSO is better than the other three algorithms.