whale optimization algorithm
در نشریات گروه عمران-
شناخت صحیح نیروی بالابرنده نقش مهمی در تحلیل پایداری سدهای وزنی دارد. لذا تخمین آن با روش های دقیق بسیار حایز اهمیت می باشد. بدین منظور یک مدل عددی از پی سد وزنی در گاوانگژو چین، به روش المان محدود انجام گرفت و بعد از شبیه سازی، مقادیر نیروی بالابرنده در موقعیت های مختلف قرارگیری زهکش حاصل گردید. نیاز به تخصص، زمان بر بودن محاسبات و تعیین دقیق شرایط اولیه مسئله در مدل های عددی سبب گردیده است که گرایش به استفاده از مدل های هوشمند گسترش پیدا کند. بدین منظور علاوه بر مدل شبکه عصبی مصنوعی مرسوم ANNبا 4 نرون ورودی، یک لایه پنهان (دارای 8 نرون) یک نرون خروجی، یک مدل جدید هیبریدی عصبی مصنوعی و الگوریتم بهینه سازی وال،ANN-WOA، توسعه داده شد. نسبت پارامترهای فاصله ردیف زهکش از بالادست، فاصله مرکز به مرکز زهکش ها از هم، قطر زهکش ها و سطح آب بالادست مخزن سد به عرض کف سد به عنوان ورودی و نیروی بالابرنده نسبی به عنوان خروجی مد نظر قرار گرفتند. نتایج تحقیق نشان داد مدل هیبریدی با مقادیر R2، RMSE و RE% به ترتیب برابر با 998/0، 021/0 و 50/3 % نسبت به مقادیر مدل شبکه عصبی مصنوعی به ترتیب برابر 995/0 ، 0261/0 و 67/4 % از قابلیت بالایی در تخمین نیروی بالابرنده برخوردار می باشد. همچنین نمودارهای چگالی داده ها و دیاگرام ویلن نشان داد که پراکندگی و توزیع احتمال داده های تخمینی با مدل هیبریدی با داده های حاصل از شبیه سازی عددی تطابق بسیار نزدیک و مشابهی دارد.
کلید واژگان: سد وزنی، نیروی بالابرنده، روش اجزا محدود، هیبرید شبکه عصبی مصنوعی، الگوریتم بهینه سازی والThe correct identification of the uplift force plays an important role in the stability analysis of gravity dams. For this purpose, a numerical model of the foundation of a gravity dam of the Guangzhao, China was made using finite element method. After simulation, the uplift force values were obtained in different positions of drainage. Require experience, the timing of calculations and the accurate determination of the boundary conditions in numerical models, have caused to the development of the tendency to use intelligent models. For this purpose, in addition to the Artificial Neural Network model (ANN) with three-layer that consists of 4 input neurons, 1 hidden layer (with 8 neurons), and 1 output neurons, a new hybrid model of Artificial Neural Network-Whale Optimization Algorithm (ANN-WOA), was developed. The values of R2, RMSE and RE% for the ANN-WOA model, were 0.998, 0.021 and 3.3%, respectively, and for the ANN model were 0.995, 0.261 and 4.67% respectively, that indicate the higher accuracy of the ANN-WOA model in the estimation of the uplift force than the ANN. In addition, the density plot box and the violin plot indicate that the point density and the probability distribution estimated data with the ANN-WOA model is very similar to that the data obtained from the numerical simulation compared with the ANN model.
Keywords: gravity dam, Uplift force, finite element method, Hybrid artificial neural network, whale optimization algorithm -
In this study, the performance of the algorithms of whale, Differential evolutionary, crow search, and Gray Wolf optimization were evaluated to operate the Golestan Dam reservoir with the objective function of meeting downstream water needs. Also, after defining the objective function and its constraints, the convergence degree of the algorithms was compared with each other and with the absolute optimal values obtained from GAMS nonlinear programming method (19.41). These values together with each algorithm optimization results were ranked using six multi-criteria decision-making methods of TOPSIS, VICOR, Linmap, Codas, ELECTRE and Simple Additive Weighting after obtaining the performance evaluation criteria of each algorithm (Reliability, reversibility, and vulnerability). Finally, integration methods (Mean, Borda, and Copland techniques) were used to evaluate the performance of decision models. The results showed that the mean responses of the gray wolf, the whale, differential evolutionary, and crow search algorithms were 1.08, 1.49, 1.29 and 1.19 times the absolute optimal response and the answers’ coefficient of variation obtained by Gray Wolf algorithm was 113.2, and 1.43 times smaller than the whale, differential evolutionary, and crow search algorithms, respectively. Moreover, all integration techniques indicated the superiority of the gray wolf algorithm. Then, the Crow search, Differential evolutionary, and whale algorithms were ranked second to fourth, respectively. On the other hand, the use of these methods in solving the problem of Golestan Dam reservoir optimization was considered appropriate due to the similarity of the results obtained from the integration techniques with the results of TOPSIS, VICOR and Linmap methods.Keywords: Optimal use of dam reservoir, Whale optimization algorithm, Differential evolutionary optimization algorithm, Crow search optimization algorithm, Gray wolf optimization algorithm
-
International Journal of Optimization in Civil Engineering, Volume:9 Issue: 2, Spring 2019, PP 213 -232Many researches have focused on the optimal design of tuned mass damper (TMD) system without the effect of soil–structure interaction (SSI), so that ignoring the effect of SSI may lead to an undesirable and unrealistic design of TMD. Furthermore, many optimization criteria have been proposed for the optinal design of the TMD system. Hence, the main aim of this study is to compare different optimization criteria for the optimal design of the TMD system considering the effects of SSI in a high–rise building. To acheive this purpose, the optimal TMD for a 40–storey shear building is firstly evaluated by expressing the objective functions in terms of the reduction of structural responses (including the displacement and acceleration) and the limitation of the scaled stroke of TMD. Then, the best optimization criterion is selected, which leads to the best performance for the vibration control of the structure. In this study, the whale optimization algorithm (WOA) is employed to optimize the parameters of the TMD system. The numerical results show that the soil type and selected objective function efficiently affect the optimal design of the TMD system.Keywords: tuned mass damper, soil–structure interaction, optimization criteria, opimal design, whale optimization algorithm, transfer function
- نتایج بر اساس تاریخ انتشار مرتب شدهاند.
- کلیدواژه مورد نظر شما تنها در فیلد کلیدواژگان مقالات جستجو شدهاست. به منظور حذف نتایج غیر مرتبط، جستجو تنها در مقالات مجلاتی انجام شده که با مجله ماخذ هم موضوع هستند.
- در صورتی که میخواهید جستجو را در همه موضوعات و با شرایط دیگر تکرار کنید به صفحه جستجوی پیشرفته مجلات مراجعه کنید.