Projects & Grants
| Development of adaptive mechanisms of evolutionary algorithms in order to solve various optimization problems. | |
|---|---|
| Project Id | SGS07/ÚVAFM/20 |
| Main solver | doc. RNDr. Petr Bujok, Ph.D. |
| Period | 1/2020 - 12/2020 |
| Provider | Specifický VŠ výzkum |
| State | finished |
| Anotation | Evolutionary algorithms (EA) are as a standard for solving of optimization tasks. These methods are inspired by various nature-ecosystems, one of the most applied EA is differential evolution (DE). Variants of DE achieve very excelent results among all optimization methods. This is caused by its simplicity, low number of parameters, and transparent structure. Despite the very good results, there is still place to enhance the DE algorithm. Weakness of EA is in necessity to set the control parameters in accordance with current optimization problem. Nevertheless, many adaptive DE variants need to set small collection of control parameters by user. In spite of that facts, there is still place to propose such algorithm, which will solve higher number of optimization tasks better than existing methods. In the proposed project, enhanced mechanisms for parameters adaptation in DE, or generally EA, will be introduced. Next area of research of the project is setting of the parameters in accordance with goal function landscape of the solved task. Further, a controlled cooperation of various EAs will be studied, because this area provided good results in previous research. As a part of the proposed project, new EAs for solving large-scale optimization tasks will be studied and developed. The main reason is that many real optimization tasks are defined by high number of parameters. The main goal of the proposed project is development and application of new evolutionary algorithms to achieve better results. In practise, there are many tasks to be optimized. Therefore, part of the project will be focused on this area of research. |
