ﺑﺮرﺳﯽ ﻗﺎﺑﻠﯿﺖ ﮐﺎرﺑﺮد اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي ﮔﺮگ ﺧﺎﮐﺴﺘﺮي در ﺗﻌﯿﯿﻦ اﺑﻌﺎد ﺑﻬﯿﻨﻪ ﻣﻘﻄﻊ ﺳﺪﻫﺎي ﺑﺘﻨﯽ وزﻧﯽ

Investigate the Applicability of Gray Wolf Optimization Algorithm in Determining the Optimal Dimensions of Concrete Dams

ﻃﺮاﺣﯽ ﺑﻬﯿﻨﻪ اﺑﻌﺎد ﺳﺪﻫﺎي ﺑﺘﻨﯽ وزﻧﯽ ﺑﺎ ﮐﺎﻫﺶ ﺳﻄﺢ ﻣﻘﻄﻊ ﺳﺪ ﺑﻪ ﮐﺎﻫﺶ ﺣﺠﻢ ﺑﺘﻦ ﻣﺼﺮﻓﯽ و ﮐﺎﻫﺶ ﻫﺰﯾﻨﻪﻫﺎي ﺳﺎﺧﺖ ﻣﯽاﻧﺠﺎﻣﺪ. ﺑﻪ دﻟﯿﻞ ﺗﻌﺪد ﻗﯿﻮدات ﺣﺎﮐﻢ ﺑﺮ ﻣﺴﺌﻠﻪ ﻫﻤﭽﻮن ﻗﯿﺪﻫﺎي ﭘﺎﯾﺪاري در ﻣﻘﺎﺑﻞ واژﮔﻮﻧﯽ و ﻟﻐﺰش ﮐﻪ ﺑﻪ ﭘﯿﭽﯿﺪه ﺷﺪن ﻓﻀﺎي ﺗﺼﻤﯿﻢ ﻣﯽاﻧﺠﺎﻣﺪ، اﺳﺘﻔﺎده از ﻣﺪلﻫﺎي ﺑﻬﯿﻨﻪﺳﺎزي ﻓﺮاﮐﺎوﺷﯽ در ﻃﺮاﺣﯽ ﺑﻬﯿﻨﻪ ﺳﺪﻫﺎي ﺑﺘﻨﯽ وزﻧﯽ ﮔﺴﺘﺮشﯾﺎﻓﺘﻪ اﺳﺖ. در اﯾﻦ ﭘﮋوﻫﺶ ﺑﺮاي ﻧﺨﺴﺘﯿﻦ ﺑﺎر ﻗﺎﺑﻠﯿﺖ اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي ﮔﺮگ ﺧﺎﮐﺴﺘﺮي در ﺣﻞ ﻣﺴﺌﻠﻪ ﻃﺮاﺣﯽ ﺑﻬﯿﻨﻪ اﺑﻌﺎد ﺳﺪ ﺑﺘﻨﯽ وزﻧﯽ ﮐﻮﯾﻨﺎ ﻣﻮردﺑﺮرﺳﯽ ﻗﺮارﮔﺮﻓﺘﻪ و اﺑﻌﺎد ﺑﻬﯿﻨﻪ اﺳﺘﺨﺮاﺟﯽ ﺑﺎ ﻧﺘﺎﯾﺞ اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي ﻫﻮش ﺗﺠﻤﻌﯽ ذرات ﻣﻘﺎﯾﺴﻪ ﺷﺪ. ﻣﻘﺎﯾﺴﻪ ﻧﺸﺎن داد ﮐﻪ اﻟﮕﻮرﯾﺘﻢ ﮔﺮگ ﺧﺎﮐﺴﺘﺮي ﺑﺎ 61/5 درﺻﺪ ﺗﻌﺪاد ﻓﺮاﺧﻮان ﮐﻢﺗﺮ ﻧﺴﺒﺖ ﺑﻪ اﻟﮕﻮرﯾﺘﻢ ﻫﻮش ﺗﺠﻤﻌﯽ ذرات، ﺗﻮاﻧﺴﺘﻪ ﻣﯿﺎﻧﮕﯿﻦ ﺗﻮاﺑﻊ ﻫﺪف ﻣﺤﺎﺳﺒﺎﺗﯽ را ﺑﻪ ﻣﯿﺰان 5/2 درﺻﺪ ﺑﻬﺒﻮد ﺑﺨﺸﺪ. ازﻧﻈﺮ ﭘﺎﯾﺪاري راهﺣﻞﻫﺎ ﻧﯿﺰ اﻧﺤﺮاف ﻣﻌﯿﺎر ﺗﻮاﺑﻊ ﻫﺪف ﻣﺤﺎﺳﺒﻪﺷﺪه ﺑﺎ اﻟﮕﻮرﯾﺘﻢ ﮔﺮگ ﺧﺎﮐﺴﺘﺮي ﻧﺴﺒﺖ ﺑﻪ اﻟﮕﻮرﯾﺘﻢ ﻫﻮش ﺗﺠﻤﻌﯽ ذرات ﮐﻢﺗﺮ اﺳﺖ. ﻫﺮﭼﻨﺪ ازﻧﻈﺮ رﺳﯿﺪن ﺑﻪ ﺟﻮاب ﺑﻬﯿﻨﻪ ﻣﻄﻠﻖ در ده ﺑﺎر اﺟﺮاي اﻟﮕﻮرﯾﺘﻢ، ﺟﻮاب ﻣﺤﺎﺳﺒﻪﺷﺪه ﺗﻮﺳﻂ اﻟﮕﻮرﯾﺘﻢ ﻫﻮش ﺗﺠﻤﻌﯽ ذرات ﺑﻪ ﻣﯿﺰان 1/6 درﺻﺪ ﺑﻬﺘﺮ اﺳﺖ. درﻣﺠﻤﻮع ﻧﺘﺎﯾﺞ ﻧﺸﺎندﻫﻨﺪه ﮐﺎرآﯾﯽ اﻟﮕﻮرﯾﺘﻢ ﮔﺮگ ﺧﺎﮐﺴﺘﺮي در ﻣﺤﺎﺳﺒﻪ اﺑﻌﺎد ﺳﺪﻫﺎي وزﻧﯽ ﺑﺘﻨﯽ از ﻧﻈﺮ دﻗﺖ و ﭘﺎﯾﺪاري ﺟﻮابﻫﺎ و ﺳﺮﻋﺖ همگرايي است

Acquiring optimal dimensions of concrete dams by reducing the cross section will reduce the volume of concrete used and consequently reduce construction costs. Due to multiplicity of constraints which leads to complexity of the decision space, meta-heuristic optimization algorithms have been increasingly used in optimal design of gravity dams. In this study, performance of Grey Wolf Optimizer was assessed in optimal design of Koyna concrete gravity dam and the results were compared to Particle Swarm Optimization. Results indicated that Grey Wolf Optimization was able to reach to a 5.2 percent lower mean optimal cost over 10 runs result with considerably a 61.5 percent smaller number of function evaluations compared to Particle Swarm Optimization and considering stability of the algorithm results obtained from Grey Wolf Optimization had smaller standard deviation than Particle Swarm Optimization. However, in 10 consecutive runs, the minimum optimal value reached using Particle Swarm Optimization was 1.6% smaller than the smallest optimal value obtained from Grey Wolf Optimization, considering overall performance of algorithms, Grey Wolf Optimization presented a promising performance having a faster and stable convergence.