پديد آورنده :
رجبي رناني، محسن
عنوان :
پيش‌بيني رفتار جمعي نهايي نوسانگرهاي مدل كوراموتو در شبكه‌ي حلقه با استفاده از رهيافت يادگيري ماشين
مقطع تحصيلي :
كارشناسي ارشد
محل تحصيل :
اصفهان : دانشگاه صنعتي اصفهان
صفحه شمار :
سيزده، 46ص. : مصور، جدول، نمودار
استاد راهنما :
فرهاد شهبازي، جلال ذهبي
توصيفگر ها :
مدل كوراموتو , يادگيري ماشين , رفتار جمعي
استاد داور :
كيوان آقابابايي ساماني، فرهاد فضيله
تاريخ ورود اطلاعات :
1400/10/18
تاريخ ويرايش اطلاعات :
1400/10/18
چكيده فارسي :
ﺑﺨﺶ ﺍﻋﻈﻢ ﺳﺎﻣﺎﻧﻪﻫﺎﯼ ﺩﯾﻨﺎﻣﯿﮑﯽ ﺭﺍ ﺳﺎﻣﺎﻧﻪﻫﺎﯾﯽ ﺗﺸﮑﯿﻞ ﻣﯽﺩﻫﻨﺪ ﮐﻪ ﺑﻪ ﺧﺎﻃﺮ ﭘﯿﭽﯿﺪﮔﯽﻫﺎﯼ ﺭﻭﺍﺑﻂ ﻭ ﺷﺒﮑﻪﯼ ﺍﺟﺰﺍ، ﺩﯾﻨﺎﻣﯿﮏ ﺁﻥﻫﺎ ﺑﺎ ﺭﻭﺵﻫﺎﯼ ﺗﺤﻠﯿﻠﯽ ﺳﻨﺘﯽ ﻗﺎﺑﻞ ﭘﯿﺶﺑﯿﻨﯽ ﻧﯿﺴﺖ ﻭ ﭘﯿﺶﺑﯿﻨﯽ ﺭﻓﺘﺎﺭ ﺍﯾﻦ ﺳﺎﻣﺎﻧﻪﻫﺎ ﯾﮏ ﻣﻮﺿﻮﻉ ﺑﻪﺭﻭﺯ ﺩﺭ ﻓﯿﺰﯾﮏ ﻣﺤﺴﻮﺏ ﻣﯽﺷﻮﺩ. ﭘﺪﯾﺪﻩﯼ ﻫﻤﮕﺎﻣﯽ ﯾﮑﯽ ﺍﺯ ﭘﺪﯾﺪﻩﻫﺎﯾﯽ ﺍﺳﺖ ﮐﻪ ﺑﻪ ﻭﻓﻮﺭ ﺩﺭ ﺍﻧﻮﺍﻉ ﻣﺨﺘﻠﻔﯽ ﺍﺯ ﺳﺎﻣﺎﻧﻪﻫﺎﯼ ﺩﯾﻨﺎﻣﮑﯽ ﻗﺎﺑﻞﻣﺸﺎﻫﺪﻩ ﺍﺳﺖ. ﻣﺪﻝ ﮐﻮﺭﺍﻣﻮﺗﻮ ﯾﮑﯽ ﺍﺯ ﺳﺎﺩﻩﺗﺮﯾﻦ ﻣﺪﻝﻫﺎﯾﯽ ﺍﺳﺖ ﮐﻪ ﺑﺮﺍﯼ ﺗﻮﺻﯿﻒ ﺍﯾﻦ ﭘﺪﯾﺪﻩ ﺍﺭﺍﺋﻪ ﺷﺪﻩ ﺍﺳﺖ ﮐﻪ ﺩﺭ ﻋﯿﻦ ﺳﺎﺩﻩ ﺑﻮﺩﻥ ﺳﺎﺯﮔﺎﺭﯼ ﺧﻮﺑﯽ ﺑﺎ ﻧﺘﺎﯾﺞ ﺗﺠﺮﺑﯽ ﺩﺍﺷﺘﻪ ﺍﺳﺖ. ﺩﺭ ﺍﯾﻦ ﭘﮋﻭﻫﺶ ﺳﻌﯽ ﺷﺪ ﺗﺎ ﺑﺎ ﺭﻭﯾﮑﺮﺩﯼ ﻣﺘﻔﺎﻭﺕ ﻧﺴﺒﺖ ﺑﻪ ﻗﺒﻞ ﺭﻓﺘﺎﺭ ﺟﻤﻌﯽ ﻧﻬﺎﯾﯽ ﻧﻮﺳﺎﻧﮕﺮﻫﺎﯼ ﮐﻮﺭﺍﻣﻮﺗﻮ ﺭﻭﯼ ﺷﺒﮑﻪﯼ ﺣﻠﻘﻪ ﺑﺎ ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﺭﻫﯿﺎﻓﺖ ﯾﺎﺩﮔﯿﺮﯼ ﻣﺎﺷﯿﻦ ﭘﯿﺶﺑﯿﻨﯽ ﺷﻮﺩ. ﺩﺭ ﺣﺎﻟﺖ ﻋﺎﺩﯼ ﺑﺮﺍﯼ ﺑﻪﺩﺳﺖ ﺁﻭﺭﺩﻥ ﺣﺎﻟﺖ ﻧﻬﺎﯾﯽ ﺍﯾﻦ ﺩﯾﻨﺎﻣﯿﮏ ﻧﯿﺎﺯ ﺩﺍﺭﯾﻢ ﺗﺎ ﻓﺮﺁﯾﻨﺪ ﺁﻥ ﺷﺒﯿﻪﺳﺎﺯﯼ ﮐﺎﻣﭙﯿﻮﺗﺮﯼ ﺷﻮﺩ ﮐﻪ ﺍﯾﻦ ﻣﺴﺌﻠﻪ ﻫﺰﯾﻨﻪﻫﺎﯼ ﻣﺤﺎﺳﺒﺎﺗﯽ ﺧﻮﺩ ﺭﺍ ﺩﺍﺭﺩ ﻭ ﺍﮔﺮ ﺑﺘﻮﺍﻧﯿﻢ ﺗﻮﺍﻧﺎﯾﯽ ﭘﯿﺶﺑﯿﻨﯽ ﺣﺎﻟﺖ ﻧﻬﺎﯾﯽ ﺍﯾﻦ ﺩﯾﻨﺎﻣﯿﮏ ﺭﺍ ﺑﻪﺩﺳﺖ ﺁﻭﺭﯾﻢ ﻗﺎﺩﺭﯾﻢ ﺑﺎ ﻫﺰﯾﻨﻪﻫﺎﯼ ﻣﺤﺎﺳﺒﺎﺗﯽ ﺧﯿﻠﯽ ﮐﻤﺘﺮ ﺣﺎﻟﺖ ﻧﻬﺎﯾﯽ ﺍﯾﻦ ﺳﺎﻣﺎﻧﻪ ﺭﺍ ﭘﯿﺪﺍ ﮐﻨﯿﻢ. ﺍﺯ ﻃﺮﻑ ﺩﯾﮕﺮ ﺍﯾﻦ ﻣﻮﺿﻮﻉ ﺍﺯ ﺍﯾﻦ ﻧﻈﺮ ﺣﺎﺋﺰ ﺍﻫﻤﯿﺖ ﺍﺳﺖ ﮐﻪ ﺍﺳﺘﻔﺎﺩﻩ ﺍﺯ ﺍﯾﻦ ﺍﻟﮕﻮﺭﯾﺘﻢﻫﺎ ﺩﺭ ﺣﻞ ﻣﺴﺎﺋﻞ ﻓﯿﺰﯾﮑﯽ ﺗﺎﺭﯾﺨﭽﻪﯼ ﺯﯾﺎﺩﯼ ﻧﺪﺍﺭﺩ ﻭ ﻣﻮﻓﻘﯿﺖ ﺩﺭ ﺍﯾﻦ ﭘﮋﻭﻫﺶ ﻣﯽﺗﻮﺍﻧﺪ ﺳﺮﺁﻏﺎﺯ ﺍﺳﺘﻔﺎﺩﻩﯼ ﺑﯿﺸﺘﺮ ﺍﺯ ﺁﻥﻫﺎ ﺩﺭ ﺣﻞ ﻣﺴﺎﺋﻞ ﺳﺨﺖﺗﺮ ﻓﯿﺰﯾﮏ ﺍﺯ ﺟﻤﻠﻪ ﻣﺴﺎﺋﻞ ﻣﺮﺑﻮﻁ ﺑﻪ ﺩﯾﻨﺎﻣﯿﮏ ﻏﯿﺮﺧﻄﯽ ﻭ ﺁﺷﻮﺏ ﺑﺎﺷﺪ.
چكيده انگليسي :
Most of the dynamic systems are systems that cannot be predicted by traditional analytical methods due to the complexity of relationships and network of their dynamic components, and predicting the behavior of these systems is an up-to-date topic in physics. The synchronization phenomenon is one of the phenomena that can be observed in abundance in different types of dynamic systems. The Kuramoto model is one of the simplest and at the same time the best models to describe this phenomenon, which has been in good agreement with the experimental results. In this study, we tried to predict the final collective behavior of Kuramoto oscillators on the ring network with a different approach than before using the machine learning approach. Normally, obtaining the final state of this dynamic requires computer simulations, which have their computational costs, and if we can obtain the ability to predict the final state of this dynamic, we can achieve the result of this system with much lower computational costs. On the other hand, it is important to note that the use of these algorithms in solving physical problems does not have much history, and success in this research could be the beginning of using more of them in solving more difficult physics problems, including problems with nonlinear .dynamics and chaos
استاد راهنما :
فرهاد شهبازي، جلال ذهبي
استاد داور :
كيوان آقابابايي ساماني، فرهاد فضيله
