Differential evolution with biological-based mutation operator
| Autor: | Raghav Yadav, Shashi Prabha |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2020 |
| Předmět: |
Mathematical optimization
Computer Networks and Communications Computer science 020209 energy Population Evolutionary algorithm 02 engineering and technology Biomaterials Operator (computer programming) 0202 electrical engineering electronic engineering information engineering Local search (optimization) education Civil and Structural Engineering Fluid Flow and Transfer Processes education.field_of_study Hemostasis business.industry Mechanical Engineering 020208 electrical & electronic engineering Metals and Alloys Electronic Optical and Magnetic Materials Rate of convergence Self-adaptive Hardware and Architecture lcsh:TA1-2040 Differential evolution Mutation (genetic algorithm) Mutation strategy business lcsh:Engineering (General). Civil engineering (General) Premature convergence |
| Zdroj: | Engineering Science and Technology, an International Journal, Vol 23, Iss 2, Pp 253-263 (2020) |
| ISSN: | 2215-0986 |
| Popis: | Evolutionary algorithms are swiftly gaining recognition where the application of the same in solving complex real-world problems is beneficial. One of the best well known Evolutionary algorithms is Differential Evolution (DE). It is considered the most powerful stochastic population-based optimization technique, inspired by the natural phenomenon of evolution to solve the real world problems. The performance of differential evolution has been promising, enabling researchers to develop a DE-variant that prevents premature convergence and enable delayed stagnation. The proposed method involves the implementation of new mutation vectors inspired from the biological phenomenon called Hemostasis that regulates the blood flow in the human body. The proposed bio-inspired mutation operator is termed as a Hemostatic operator that results in more promising solutions. It enhances diversity by means of good vectors during earlier stages thereby avoiding stagnation in the later stages. Actually, the process to avoid trapping in local search space pushes the proposed algorithm to perform globally in an effective manner. Specifically, it accelerates the convergence rate to enhance the search capability towards global optima and assist differential evolution to address the problem of stagnation. It has been compared with other state-of-the-art optimization algorithms on COCO (Comparing Continuous Optimizers). The simulation results indicate that the proposed Hemostasis based mutation strategy outperforms most of the state-of-the-art differential evolution variants and few population-based optimization algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|