Multi‐objective firefly algorithm with multi‐strategy integration

Abstract

In the optimization process of multi-objective firefly algorithm, population is easy to fall into local optimum, which leads to poor population distribution and convergence. In order to solve this problem, this article proposes a multi-objective firefly algorithm with multi-strategy integration (MOFA-MSI). First, in order to improve the distribution of population, MOFA-MSI proposes a cloning strategy, which calculates the distribution degree of individuals in population, clones them according to their distribution degree, and local mutation in the cloned individual produce a new population with good distribution. Then, in order to maintain the convergence of population, a position updating strategy based on non-dominated sorting is proposed. The new population after local mutation are performed by non-dominated sorting, and the fireflies with higher rank guide the fireflies with lower rank to fly, and then new population are generated by global mutation after position updated. Finally, the greedy strategy is adopted to select solutions with better distribution and convergence and store them in external files. In the experimental part, different types of test problems are used to test the performance of each algorithm, and MOFA-MSI is compared with three classical and four new multi-objective evolutionary algorithms. The results show that MOFA-MSI is superior to other seven algorithms in terms of the distribution and convergence of population.

Task Scheduling in Cloud Computing Environment Using Advanced Phasmatodea Population Evolution Algorithms

“…On a soft real-time single processor system, Fahmy developed a fuzzy method for aperiodic work scheduling. The total throughput time of the task is lowered by assessing the priority of the work that is now running and altering the priority of the job in the queue, and the technique is utilized in a multiobjective algorithm [46]. Zhou et al [47] proposed a heterogeneous earliest completion time algorithm based on fuzzy dominance sorting in the Infrastructure as a Service (IaaS) workflow, which greatly improved the running speed.…”

Tri-training algorithm based on cross entropy and K-nearest neighbors for network intrusion detection

“…In the past machine learning process, the relative entropy KL D is mainly used to determine the difference between the practical probability distribution ( ) P X and the predicted probability distribution ( ) Q X . From Eq (6), it is clear that the difference between ( ) P X and ( ) Q X can be determined by cross entropy, and it is more convenient to calculate the differexnce between two probability distributions by using cross entropy than relative entropy.…”

“…Traditional NID techniques achieve intrusion detection by comparing attacks identified in the feature code database, but this method has a high leakage rate and lag [1]. Recently, with the rapid development of artificial intelligence [2][3][4][5] and machine learning technologies [6][7][8], machine learning-based NID methods are gradually becoming a research hotspot.…”

Comparison of metaheuristic optimization algorithms for numerical solutions of optimal control problems

“…7 This algorithm is applied in many different fields such as GA. These algorithms can be used in many different fields such as extracting features in image processing, [37][38][39][40] engineering problems, 13,[41][42][43][44] solving NP hard problems, [45][46][47] classification, clustering 48 and health. 36,49 Such algorithms provide successful results in cases where the objective function is discontinuous and in limited optimization operations.…”