استفاده از تكنيك‌هاي محاسبات نرم جهت برنامه‌ريزي تخصيص منابع آب در زمان خشكسالي

The Use of Soft Computing Techniques for Irrigation Scheduling During Drought Episode

ﺑﺨﺶ ﮐﺸﺎورزي ﻋﻤﺪه ﻣﺼﺮف ﮐﻨﻨﺪه ﻣﻨﺎﺑﻊ آب در ﮐﺸﻮر ﻣﯽﺑﺎﺷﺪ، ﻟﺬا اﺗﺨﺎذ ﺗﺼﻤﯿﻤﺎت ﻣﻨﺎﺳﺐ در ﺑﺮﻧﺎﻣﻪرﯾﺰي و ﺗﺨﺼﯿﺺ ﻣﻨﺎﺑﻊ آب در اﯾﻦ ﺑﺨﺶ ﮐﻤﮏ ﺷﺎﯾﺎﻧﯽ در ﺟﻬﺖ ﻣﺪﯾﺮﯾﺖ ﮐﺎراي اﯾﻦ ﻣﻨﺎﺑﻊ ﻣﯽﮐﻨﺪ. ﻫﺪف ﺗﺤﻘﯿﻖ ﺣﺎﺿﺮ آﻧﺴﺖ ﺗﺎ ﺑﺎ اﺳﺘﻔﺎده از اﺑﺰارﻫﺎي ﻣﺤﺎﺳﺒﺎت ﻧﺮم ﻫﻤﺎﻧﻨﺪ اﻟﮕﻮرﯾﺘﻢ ﺑﻬﯿﻨﻪﺳﺎزي ازدﺣﺎم ذرات )PSO( و اﻟﮕﻮرﯾﺘﻢ ژﻧﺘﯿﮏ )GA(، ﻣﺪﻟﺴﺎزي ﻣﻘﺎدﯾﺮ رﻫﺎﺳﺎزي ﺑﻬﯿﻨﻪ از ﻣﺨﺰن و ﺑﺮﻧﺎﻣﻪرﯾﺰي آﺑﯿﺎري در ﺷﺒﮑﻪﻫﺎي ﮐﺸﺎورزي ﭘﺎﯾﯿﻦدﺳﺖ ﺳﺪ زاﯾﻨﺪه¬رود، ﺗﺪوﯾﻦ ﮔﺮدد. در اﯾﻦ راﺳﺘﺎ ﺗﻘﻮﯾﻢ ﮐﺸﺖ ﻣﺤﺼﻮﻻت، ﮐﻞ آ ب ﻣﻮﺟﻮد و زﻣﯿﻦ ﻗﺎﺑﻞ ﮐﺸﺖ در ﺑﺨﺶ ﮐﺸﺎورزي و ﻣﺤﺪودﯾﺖﻫﺎي ﻧﯿﺎز آﺑﯽ ﻣﺘﻐﯿﺮ ﻣﺤﺼﻮﻻت از ﺟﻤﻠﻪ ﻣﻬﻤﺘﺮﯾﻦ ﻗﯿﻮد ﻏﯿﺮ ﺧﻄﯽ ﺗﺤﻘﯿﻖ ﺣﺎﺿﺮ ﻣﯽﺑﺎﺷﻨﺪ. ﻧﺘﺎﯾﺞ ﺣﺎﮐﯽ از آن ﺑﻮد ﮐﻪ ﻣﺪﻟﺴﺎزي ﯾﮑﭙﺎرﭼﻪ PSO ﺑﺎ ﺗﻮزﯾﻊ ﺑﻬﺘﺮ ﮐﻤﺒﻮدﻫﺎي آﺑﯽ ﺑﯿﻦ ﻣﺮاﺣﻞ ﻣﺨﺘﻠﻒ رﺷﺪ ﻣﺤﺼﻮﻻت، ﺗﻮاﻧﺴﺖ ﺳﻮد ﺳﯿﺴﺘﻢ را در ﻣﻘﺎﯾﺴﻪ ﺑﺎ ﻣﻘﺎدﯾﺮ ﺗﺨﺼﯿﺺ ﺑﻪ روش ﺳﻨﺘﯽ )وﻧﺶﺑﻨﺪي( ﺑﻪ ﻃﻮر ﻗﺎﺑﻞ ﺗﻮﺟﻬﯽ اﻓﺰاﯾﺶ دﻫﺪ. ﻫﻤﭽﻨﯿﻦ ﺑﺎ ﻣﻼﺣﻈﻪ زﻣﺎن رﺳﯿﺪن ﺑﻪ ﻫﻤﮕﺮاﯾﯽ و ﻧﯿﺰ ﮐﻞ ﺳﻮد، ﻣﺪل PSO ﻋﻤﻠﮑﺮد ﺑﻬﺘﺮي ﻧﺴﺒﺖ ﺑﻪ GA داﺷﺘﻪ اﺳﺖ. در ﻧﻬﺎﯾﺖ اﺳﺘﻔﺎده از ﺗﮑﻨﯿﮏﻫﺎي ﻣﺤﺎﺳﺒﺎت ﻧﺮم در ﺑﺮﻧﺎﻣﻪرﯾﺰي آﺑﯿﺎري، ﻣﯽﺗﻮاﻧﺪ اﻟﮕﻮﻫﺎي ﻣﻮﺛﺮي ﺑﺮاي ﺗﺨﺼﯿﺺ ﻣﻨﺎﺑﻊ آب ﮐﺸﺎورزي در ﻣﻨﺎﻃﻖ ﺧﺸﮏ ﺑﺎ ﻣﻨﺎﺑﻊ آب ﻣﺤﺪود ﻓﺮاﻫﻢ ﮐﻨﺪ.

Agricultural sector is the main water consumer in our country. So the appropriate decisions for irrigation scheduling and its optimal allocation is of great importance for an efficient water management. The aim of the present study is to employ some soft computing techniques, such as the particle swarm optimization (PSO) and genetic algorithm (GA), and to determine optimal irrigation scheduling as well as reservoir release for agricultural networks located at downstream of Zayandeh-Rud dam. In this regard, the crop calendar, total amount of available water as well as arable land in agricultural sector, the amount of water available at the beginning of water year and crop water requirements are the most important non-linear constraints of current research. The results showed the integrated PSO modeling with better distribution of water shortages among different crop growth stages could significantly increase the net profit of system while compared to those of traditional irrigation systems. Regarding the time of reaching the convergence as well as total attainable benefit, the PSO has slightly outperformed the GA. Consequently, application of soft computing techniques in irrigation scheduling will provide effective water allocation patterns to save more water in an arid region with limited water resources.