Opposition-based ensemble micro-differential evolution

Abstract

Differential evolution (DE) algorithm with a small population size is called Micro-DE (MDE). A small population size decreases the computational complexity but also reduces the exploration ability of DE by limiting the population diversity. In this paper, we propose the idea of combining ensemble mutation scheme selection and opposition-based learning concepts to enhance the diversity of population in MDE at mutation and selection stages. The proposed algorithm enhances the diversity of population by generating a random mutation scale factor per individual and per dimension, randomly assigning a mutation scheme to each individual in each generation, and diversifying individuals selection using opposition-based learning. This approach is easy to implement and does not require the setting of mutation scheme selection and mutation scale factor. Experimental results are conducted for a variety of objective functions with low and high dimensionality on the CEC Black-Box Optimization Benchmarking 2015 (CEC-BBOB 2015). The results show superior performance of the proposed algorithm compared to the other micro-DE algorithms.

Metaheuristics in the Balance: A Survey on Memory‐Saving Approaches for Platforms with Seriously Limited Resources

“…Since then, several micropopulation Evolutionary Algorithms (μ EAs) followed, such as, e.g., [83]. After understanding the potential of a classic micro-Diferential Evolution (μ DE) over large-scale problems, several μ DE schemes, self-adaptive variants (such as μ JADE), and hybrid memetic alterations were proposed [75,78,[84][85][86][87][88][89][90][91]. Analogously, Swarm Intelligence algorithms have been shown to have similar advantages when run with micropopulations.…”

A vectorized bimodal distribution based micro differential evolution algorithm (VB-mDE)

“…In MDEVM, by randomizing and vectorizing the scale factor, the diversity of the population can be increased, thereby alleviating the problems of premature and stagnation. Later, by employing ensemble mutation and oppositional learning strategies in MDEVM, the ensemble mDE (EMDE) [14] and oppositional ensemble mDE (OEMDE) [18] are presented, resectively. In the mDE algorithms [13,14,18], the parameter or operator adjustment strategies are devised based on random uniform distribution, and thus the performance is limited.…”

“…The existing mDE algorithms may be summarized in Table 1. [11] Local search µDSDE [12] Directional local search; reinitialize the worse individuals deBILS [20] Best improvement local search Adjustment of parameters or operators MDEVM [13] Vectorized random scale factor EMDE [14] Vectorized random scale factor; ensemble five mutation operators OEMDE [18] Vectorized random scale factor; ensemble five mutation operators; opposition-based learning µJADE [6] Advanced adaption of scale factor and crossover rate; 'currentby-rand-to-pbest' mutation; perturbation strategy; restart strategy Applications of mDE mODE [8] Image thresholding problem deBILS [20] Topological active net optimization problems…”

“…The performances of VB-mDE/rand/1 and VB-mDE/current-to-best/1 are further compared with other four mDE algorithms, i.e., DESP [9], EMDE [14], OEMDE [18] and µJADE [6]. DESP introduces the disturbance with adaptive step size to differential mutation strategy.…”