A Statistical Analyzing Approach for Quantum Evolutionary Algorithms

This paper proposes a novel reinitialization /guidance operator for q-individuals in Quantum Evolutionary Algorithms (QEA) called Statistical Analyzing Reinitialization Quantum Evolutionary Algorithm (SARQEA). Evolutionary algorithms suffer from trapping in local optima and QEA is not an exception. In QEA, after convergence, the q-bits in qindividuals converge to true states of [0 1] or [1 0]. Trapping in local optima, the q-individuals have no chance to explore the search space. In order to improve the exploration ability of QEA and help the algorithm to escape from local optima, this paper proposes a novel reinitialization operator. In SARQEA algorithm, at first the convergence of the population is examined. If the population is converged, in the second step, using the statistical information gathered from previous searches the q-individuals are reinitialized. The new values of reinitialized q-individuals are based on the gathered information from previous searches. Several experimental results on Knapsack, Trap and some numerical function optimization algorithms are performed and the results show better performance for the proposed algorithm.