ECPO: An enhanced Crested Porcupine Optimizer

The newly published Crested Porcupine Optimizer in 2024: The A new nature inspired metaheuristic algorithm has some problems that need further modification. The CPO algorithm is divided into four stages. The defense strategies in the first and second stages (visual and sound means) are shown in the global search stage, and the defense strategies in the third and fourth stages (odor and body attacks) are shown in the local optimization stage. Crested Porcupine Optimizer: The A new nature-inspired metaheuristic algorithm still belongs to the type of meta-heuristic algorithm, so it can not escape the problem of balancing the global search ability and the fast and accurate convergence of the local optimization while avoiding the local optima. In the second stage, the problem of excessive randomness appears, which will reduce the global search ability. The local convergence ability in the chapter third stage needs to get the corresponding good sum improvement, and the result accuracy of the local rapid convergence in the fourth stage can be further improved. However, after the verification of the test data, it is necessary to adopt a new improvement idea to improve the accuracy of the algorithm, rather than simply modifying a certain step. In the initialization stage, this paper adopts the algorithm of Tent-Logistic-Cosine chaotic mapping to handle the problem of population initialization, deleted the cyclic population reduction technology, and introduces a new opposition-based learning strategy and a new dynamic updating mechanism (parameter dynamic updating strategy, sine-cosine composite perturbation factor, Tent-chios and Cauchy mutation), and changed the idea of dynamic adaptive regulation and preliminary screening. But the four stages of CPO defense remain unchanged. Finally, a classical Levy cuckoo random walk model is added to solve the problem of local optimization, with the aim to jump out of the local optimum.