A Project Portfolio Optimization Method for Distribution Network Planning Based on Improved NSGA-II

Selecting the annual investment project portfolio from the distribution network planning project library under the constraint of investment scale is currently a focus and difficulty for planners. The balance of multiple technical and economic indices is the key to the optimization process, often relying on the experience of planners. The paper initially examines the key evaluation indexes and constraints for the comprehensive planning of the distribution network, and develops a quantitative evaluation model for the optimal selection of project portfolios in distribution network planning. Subsequently, it proposes the application of NSGA-2 for model resolution and introduces variable dynamic penalty terms that reflect varying planning preferences, ensuring compliance with multiple technical criteria and investment scale constraints. During the solution process, the normal distribution crossover (NDX) operator is employed to enhance the crossover process and identify the global optimum. The accuracy of the model developed in this article is validated through a practical case study in a certain province and city.