Multi-objective Optimization of Electric Vehicle Orderly Charging Based on Variable Preprocessing and MOSO

In order to address the issue that the current CPLEX solver is unable to find the Pareto optimal solution and that intelligent algorithms perform poorly when solving the orderly charging optimization model, firstly, based on decision variable preprocessing technology, the decision variable was changed from the charging state to the charging scheme number. Next, the schedulable charging time period was restricted to reduce the variable decision space, and from there, the best preprocessing method was chosen. Secondly, the snake optimization (SO) algorithm was improved based on the fish aggregation device effects (FADs) and adapted into a multi-objective snake optimization (MOSO) algorithm to solve the multi-objective orderly charging model with the minimum net load variance of the microgrid, operating cost of the microgrid and charging cost of electric vehicles. Finally, the optimal compromise solution was selected from Pareto frontier based on TOPSIS. The numerical examples show that the variable preprocessing technology and the improved snake optimization (ISO) algorithm reduce the dimension of decision variables, eliminate the penalty term, improve the solution accuracy, and execute single objective optimization with CPLEX-like performance. When using MOSO, CPLEX linear weighted multi-objective optimization’s weight dead zone problem may be avoided, and rich Pareto solutions that take into consideration the interests of several parties can be obtained.