Objective Comparison and Selection in Mono- and Multi-Objective Evolutionary Neurocontrollers
| Autor: | Howard M. Schwartz, Ian Showalter |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Mathematical optimization
Computer science 010102 general mathematics Binary number 02 engineering and technology Auxiliary function 01 natural sciences Multi-objective optimization Hebbian theory 0202 electrical engineering electronic engineering information engineering Robot 020201 artificial intelligence & image processing 0101 mathematics Selection (genetic algorithm) |
| Zdroj: | SSCI |
| DOI: | 10.1109/ssci47803.2020.9308403 |
| Popis: | Often in multi-objective problems, several elemental objectives are combined into compound objectives by using auxiliary equations to reduce these problems to just one or two objectives. Reducing the number of objectives simplifies the problem into a more easily optimized mono-objective problem, or for multi-objective problems, reduces the Pareto front to a few dimensions for easy analysis. Here, multi-objective evolutionary neurocontrollers with both compound and elemental objectives are compared to a mono-objective evolutionary neurocontroller. The goal of this research is to compare the effectiveness of individual elemental and compound objective effectiveness, and not directly compare mono- and multi-objectivity. The effectiveness of each of the objectives is determined through a series of experiments using a previously demonstrated Lamarckian-inherited neuromodulated evolutionary neurocontroller. The evolved neurocontrollers operate a simulated vehicle pursuing a basic evader vehicle in the pursuit-evasion game. Both vehicles are subject to the effects of mass and drag. It is shown that under certain circumstances, binary objectives can be unsuitable choices as objectives, and that it can be more effective to use multi-objective solutions than to combine elemental objective problems into mono-objective problems by auxiliary functions. It is also shown that the obvious choice of objective may not be the most effective choice. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|