Remora optimization algorithm

摘要

In this paper, Remora Optimization Algorithm (ROA) is proposed, which is a new bionics-based, natural-inspired, and meta-heuristic algorithm. The inspiration for ROA is mainly due to the parasitic behavior of remora. Different locations are updated in different hosts: In some large hosts, remora feeds on the host's ectoparasites or wreckage and evades natural enemies, for example in the case of giant whales. In some small hosts, remora follows the host to move to the bait-rich area to prey, taking the fast-moving swordfish as an example. In the case of these two update methods, remora also makes some judges based on experience. If it takes the initiative to prey, it updates the host, makes a global update. If it eat around the host, remora does not change the host, and continues to local update. This algorithm is more inclined to provide a new idea for memetic algorithm, because the host in ROA can be reasonably replaced, such as ships, turtles, etc. The above dynamic mode and behavior are simulated mathematically and the validity of the ROA is tested with 29 benchmark questions and 5 actual engineering questions. Parallel comparisons are made with 10 other natural heuristics. The statistical results and comparisons show that ROA provides a very promising prospect and a strong competitive ability compared to other state-ofthe-art heuristic techniques.