摘要：Shuffled leap frog algorithm (SFLA) is a new nature-inspired intelligent algorithm, which uses the whole update and evaluation strategy on solutions. For solving multi-dimension function optimization problems, this strategy will deteriorate the convergence speed and the quality of solution of algorithm due to interference phenomena among dimensions. To overcome this shortage, a dimension by dimension improvement based on SFLA is proposed. The proposed strategy combines an updated value of one dimension with values of other dimensions into a new solution, and that whose updated value can improve the solution will be accepted greedily. Further, a new individual update formula is designed to learn experiences both from the global best and the local best solution simultaneously. Meanwhile, they also reveal the modified algorithm is competitive for continuous function optimization problems compared with other improved algorithms.
关键词：shuffled leap frog algorithm; dimension by dimension; multi-dimensional function optimization
