|
Eagle strategy is a search strategy for solving nonlinear optimization problems, and this strategy was developed by Xin-she Yang and Suash Deb, based on the foraging behaviour of eagle species such as golden eagles.〔X. S. Yang and S. Deb, Eagle strategy using Levy walk and firefly algorithms for stochastic optimization, in: Nature Inspired Cooperative Strategies for Optimization (NICSO 2010)} (Eds. J. R. Gonzalez et al.), Vol. 284, 101-111 (2010).〕 In optimization, a common strategy is to search for the optimal solution starting from a set of initial guess solutions (either random and educated guess). In the case when the cost functions are multimodal with multiple local best solutions, the final solutions may heavily depend on the initial starting solutions. In order to minimize such dependence on initial random solutions, most modern algorithms, especially metaheuristic algorithms, are able to escape local optima by using some sophisticated random techniques. However, most of these algorithms are one-stage type; that is, once initialization is done, the search process continues until an algorithm stops. Running an algorithm many times from different initial solutions may occasionally improve the overall performance on average. Eagle strategy improves this by using an iterative, interacting two-stage strategy to enhance the search efficiency by escaping the local optima and use initial solutions in different regions. It uses a slow search stage and a fast stage to simulate an eagle searching for prey tends to search on a large area and then quickly switches to a rapid chasing phase once a prey is in sight.〔M. W. Collopy, Foraging behavior and success of golden eagles, The Auk, Vol. 100, 747-749 (1983).〕 In optimization, it uses a coarse search stage on a larger area in a search space in combination with an intensive faster search algorithm in the neighbourhood of promising solutions. Two stages interchanges and proceed iteratively. As there are two stages in the strategy, each stage can employ different algorithms. For example, differential evolution can be used within eagle strategy.〔X. S. Yang and S. Deb, Two-stage eagle strategy with differential evolution, Int. J. Bio-Inspired Computation, Vol. 4(1), pp. 1-5 (2012). http://arxiv.org/abs/1203.6586〕 Studies show that such a combination is better than any of its components.〔A. H. Gandomi, X. S. Yang, S. Talatahari, S. Deb, Coupled eagle strategy and differential evolution for unconstrained and constrained global optimization, Computers & Mathematics with Applications, Vol. 63(1), 191-200 (2012). http://dx.doi.org/10.1016/j.camwa.2011.11.010〕 In the simplest case, when the first stage does not use any algorithm (just initialization), it essentially degenerates into a hill-climbing with random restart. However, this strategy could be potentially much more powerful if a good combination of different algorithms is used. == References == 抄文引用元・出典: フリー百科事典『 ウィキペディア（Wikipedia）』 ■ウィキペディアで「Eagle strategy」の詳細全文を読む スポンサード リンク
|