于家根, 劉正江, 卜仁祥, 高孝日, 李偉峰
(大連海事大學 航海學院, 遼寧 大連 116026)
基于擬態(tài)物理學優(yōu)化算法的船舶轉(zhuǎn)向避碰決策
于家根, 劉正江, 卜仁祥, 高孝日, 李偉峰
(大連海事大學 航海學院, 遼寧 大連 116026)
針對多船會遇態(tài)勢下的船舶避碰決策難題,提出一種基于擬態(tài)物理學優(yōu)化算法的船舶轉(zhuǎn)向避碰決策方法。該算法將《國際海上避碰規(guī)則》相關(guān)條款作為約束條件限定問題的可行域空間,同時考慮基于最近會遇距離和航程損失的船舶避碰目標函數(shù),通過迭代進化獲取全局范圍內(nèi)的最優(yōu)解。仿真結(jié)果表明:將擬態(tài)物理學優(yōu)化算法應(yīng)用于船舶轉(zhuǎn)向避碰決策中是可行、有效的,能為船舶駕駛員提供決策支持。
水路運輸;擬態(tài)物理學優(yōu)化算法;多船會遇;轉(zhuǎn)向避碰決策
Abstract: The Artificial Physics Optimization (APO) algorithm with the feasible space constrained by The Rules from International Regulations for Preventing Collisions at Sea is introduced into the collision avoidance decision-making in multi-ship encounter situations. The objective function is constructed on the basis of the Distance of the Closest Point of Approach (DCPA) and the voyage losses, and solved through iterative evolution to obtain the optimal solution. The simulation of the process shows that APO algorithm works effectively in determining the course alteration for avoiding collision, therefore, can help marine officers in navigation practice.
Keywords: waterway transportation; APO algorithm; multi-ship encounter; course alteration for collision avoidance
在海上實踐中,船舶避碰決策的制定主要依賴于駕駛員的專業(yè)知識及其駕駛經(jīng)驗,轉(zhuǎn)向避碰是最常見的行動方案。然而,當船舶處于復雜的多船會遇態(tài)勢時,駕駛員很難快速給出最佳避碰決策方案,且一旦作出錯誤決策將導致事故發(fā)生。對此,一些學者嘗試將進化算法、遺傳算法、蟻群算法和粒子群算法等人工智能優(yōu)化算法應(yīng)用于船舶避碰決策研究中,已取得一定的成果。[1-4]
擬態(tài)物理學優(yōu)化算法(Artificial Physics Optimization Algorithm,APO Algorithm)[5-6]是一種基于種群的啟發(fā)式隨機搜索算法,同粒子群算法相比,其具有種群多樣性好、搜索效率高的優(yōu)點。這里考慮多船會遇態(tài)勢下的轉(zhuǎn)向避碰決策,將避碰規(guī)則作為行動可行域的約束條件,利用擬態(tài)物理學優(yōu)化算法,從可行域的空間中求出目標函數(shù)極小值,進而得到轉(zhuǎn)向避碰決策。……