哈尔滨工程大学学报
哈爾濱工程大學學報
합이빈공정대학학보
JOURNAL OF HARBIN ENGINEERING UNIVERSITY
2013年
10期
1214-1220
,共7页
李敬花%胡载萍%吕慧超%孙淼
李敬花%鬍載萍%呂慧超%孫淼
리경화%호재평%려혜초%손묘
海工装备项目%多项目调度%多资源约束%模拟退火分层遗传算法
海工裝備項目%多項目調度%多資源約束%模擬退火分層遺傳算法
해공장비항목%다항목조도%다자원약속%모의퇴화분층유전산법
offshore equipment project%multi-project scheduling%multi-resource constrained%simulated annealing hierarchical genetic algorithm
为进一步缩短海工装备项目建造工期,建立了多资源约束下海工装备多项目调度的问题模型,并提出了一种基于模拟退火分层遗传算法的求解方法。该方法首先将遗传算法分为高层和低层,在低层遗传算法中设置多个特性差异较大的子种群,避免单种群进化过程中出现的过早收敛现象;然后在分层遗传算法中融入模拟退火思想,通过对交叉/变异算子及交叉/变异后个体进行模拟退火操作,克服遗传算法局部寻优能力较差的缺陷;针对研究问题的特殊性,算法对种群进行了特殊的初始化及解码操作,在保证种群多样性的同时,避免了非法个体的产生。最后通过具体算例验证了算法的可行性和有效性。
為進一步縮短海工裝備項目建造工期,建立瞭多資源約束下海工裝備多項目調度的問題模型,併提齣瞭一種基于模擬退火分層遺傳算法的求解方法。該方法首先將遺傳算法分為高層和低層,在低層遺傳算法中設置多箇特性差異較大的子種群,避免單種群進化過程中齣現的過早收斂現象;然後在分層遺傳算法中融入模擬退火思想,通過對交扠/變異算子及交扠/變異後箇體進行模擬退火操作,剋服遺傳算法跼部尋優能力較差的缺陷;針對研究問題的特殊性,算法對種群進行瞭特殊的初始化及解碼操作,在保證種群多樣性的同時,避免瞭非法箇體的產生。最後通過具體算例驗證瞭算法的可行性和有效性。
위진일보축단해공장비항목건조공기,건립료다자원약속하해공장비다항목조도적문제모형,병제출료일충기우모의퇴화분층유전산법적구해방법。해방법수선장유전산법분위고층화저층,재저층유전산법중설치다개특성차이교대적자충군,피면단충군진화과정중출현적과조수렴현상;연후재분층유전산법중융입모의퇴화사상,통과대교차/변이산자급교차/변이후개체진행모의퇴화조작,극복유전산법국부심우능력교차적결함;침대연구문제적특수성,산법대충군진행료특수적초시화급해마조작,재보증충군다양성적동시,피면료비법개체적산생。최후통과구체산례험증료산법적가행성화유효성。
To further shorten the construction period of offshore equipment projects, a model on multi-resource con-strained offshore equipment multi-project scheduling was established, and a solution based on simulated annealing hierarchical genetic algorithm ( SAHGA) was proposed. In the solution, the genetic algorithm was divided into high hierarchy and low hierarchy. In order to avoid the premature convergence phenomenon in the single population evo-lution process, a number of sub-populations with large characteristic difference were set in the low-hierarchy genetic algorithm;then, the consideration on simulated annealing was added into the hierarchical genetic algorithm, by the simulated annealing operations conducted for the crossover/mutation operators and individuals after crossover/muta-tion, the defect of inferior local optimal solution in the genetic algorithm was overcome;aiming at the speciality of the matter in discussion, in the algorithm, special initialization and decoding operation were conducted for the pop-ulation. At the same time of assuring the diversity of the population, the generation of illegal individuals was avoi-ded. Finally, an example was given to validate the feasibility and effectiveness of the algorithm.