Many-objective Optimization Using Domain Knowledge of Soil and Water Conservation Practices

Abstract

The allocation of soil and water conservation practices is a many-objective optimization problem. However, the existing multi-objective evolutionary algorithms have some defects, which limit the application in many-objective optimization problems for high-dimensional searching space. The selection and allocation of soil and water conservation practices have the unique domain knowledge. By using domain knowledge, the complexity of many-objective optimization problems with high dimension and non-linearity can be simplified to a certain extent, and the optimization cycle can be shortened. Due to the lack of research on the incorporation of evolutionary algorithm with domain knowledge, this study takes the optimization of water and soil conservation practices as an example to explore the role of domain knowledge in a many-objective evolutionary algorithm. The results show that the incorporation of domain knowledge can play a positive role in the performance of evolutionary algorithm, and can improve the quality and quantity of the solutions.