Quantum inspired evolutionary algorithm to improve parameters of neural models on example of polish electricity power exchange
Jerzy Tchórzewski, Dariusz Ruciński · 2016
Paper contains selected results of theoretical and practical research concerning the possibility of creating evolutionary algorithms inspired by quantum information technology to improve the performance of neural models. Particular focus was on calculating the quantum available for use in quantum evolutionary algorithms. It is noted that the parameters of the artificial neural network, especially the weight matrix can be improved by evolutionary algorithms. It turns out that the introduction of solutions in the field of quantum computing to evolutionary algorithms, including the creation of quantum initial population, quantum operators (crossover and mutation), and quantum selection greatly improves the accuracy of modeling, which has been verified by the examples of figures on the Polish Power Exchange Electricity. New method of creating quantum mixed numbers is also proposed.