Improving Performance of Genetic Algorithm in Circuit Optimization

This work is devoted to optimizing electronic circuits by modifying the genetic algorithm (GA) based on the previously developed idea of generalized circuit optimization. In this case, it is possible to overcome one of the main disadvantages of GA, namely, premature convergence to local minima, and significantly improve the quality of minimization. In this case, the updated GA generates a set of populations determined by different generalized optimization strategies. A control vector introduced within the generalized optimization framework determines fitness functions and chromosome parameters, such as length and structure. The control vector specifies the number of independent variables of the optimization problem and is the main element for selecting the most promising optimization strategies. Complex strategies obtained by combining different basic strategies can improve the accuracy of the optimization problem and reduce the number of required generations. As a result, it is possible to improve the quality of the resulting solution and significantly reduce the computer time needed to optimize the electronic circuit.