為了讓人工蜂群演算法加以優化,本研究對人工蜂群演算法陷入區域最佳解與搜索不精確等問題,提出二次跳躍與粒子蜂群策略改良人工蜂群演算法,二次跳躍策略包含廣域探索的一次跳躍與深度開發的二次跳躍,改良人工蜂群演算法工蜂的部分,一次跳躍可以使蜂群大幅度的搜索,讓移動失敗的蜂群,透過相對位置的映射,飛行到新的區域進行探索,二次跳躍是小幅度的開發,目的是將差勁的蜜蜂,因為全域最佳解的資訊讓蜜蜂在豐富的食物源附近進行搜索,二次跳躍策略可以改善人工蜂群演算法的移動,增加演算法的多樣性與搜索範圍;粒子蜂群策略是運用粒子群演算法方向性的概念,改良人工蜂群演算法的觀察蜂,利用全域最佳解與鄰近蜂群兩者所產生的差異向量進行引導,使觀察蜂吸收兩者的資訊,改善搜索的移動,增加演算法的開發能力。
To optimize the artificial bee colony algorithm(ABC), this study proposed double skips and PSO’ bee strategy for modifications regarding problems, such as local optimization and inexact search of the algorithm. The double skips is consisted of the single-skip global exploration and the double-skips of in-depth exploitation. In terms of modifying employee bees of ABC, single skip can result in significant jump search of the bee colony to enable bee colony of moving failure to fly to new area for exploration by mapping of relative positions. Double skips is the exploitation of small scale to allow bees of poorer capabilities to search in area rich in foods according to globally optimized information. The double skips strategy can improve the mobility of the artificial bee colony, increase algorithm diversity and widen search range. The PSO’ bee strategy uses PSO directional concepts to improve onlooker bees of ABC using the difference gap between global optimization and neighboring bee colonies for guidance. As a result, the onlooker bees can absorb information from the two sources to improve search moves and increase the exploitation capability of algorithms.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。