A new hyper-heuristic based on ant lion optimizer and Tabu search algorithm for replica management in cloud environment

Information can be shared across the Internet using cloud computing, a powerful paradigm for meeting the needs of individuals and organizations. To minimize access time and maximize load balancing for data nodes (DNs), a dynamic data replication algorithm is necessary. Even so, few of the existing algorithms consider each objective holistically during replication. An improved ant lion optimizer (ALO) algorithm and a fuzzy system are used in this paper to determine dynamically the number of replicas and the DNs for replication. Further, it balances the trade-offs among different objectives (e.g., service time, system availability, load, and monetary cost). The ALO algorithm has been widely applied to solve complex optimization problems due to its simplicity in implementation. However, ALO has premature convergence and can thus easily get trapped into the local optimum solution. In this paper, to overcome the shortcomings of ALO by balancing exploration and exploitation, a hybrid ant lion optimizer with Tabu search algorithm (ALO-Tabu) is proposed. There are several improvements of the ALO, in which the appropriate solutions are selected for the initial population based on chaotic maps (CMs) and opposition-based learning (OBL) strategies. On the other hand, there are many CMs, OBLs, and random walk strategies that make it difficult to select the best one for optimization. Generally, they are selected manually, which is time-consuming. As a result, this paper presents a hyper-heuristic ALO (HH-ALO-Tabu) that automatically chooses CMs, OBLs, and random walk strategies depending on the differential evolution (DE) algorithm. Based on 20 well-known test functions, the experiment results and statistical tests show that HH-ALO-Tabu can solve optimization problems effectively.