pso algorithm

Considering the effects of building dams in downstream basins, the need to investigate their condition and effects is one of the important issues. In this study, the optimal form of concrete double-arched dams was investigated under various interactions against earthquakes.

The volume of concrete used was considered as the goal function of the optimization problem and the design variables were the geometric parameters of the dam. The twenty geometrical parameters were investigated. First, the dam-water-foundation system was simulated using the finite element method. Then the optimization was done using the particle swarm optimization (PSO).

To check the performance of the method used to optimize arch dams, the Moropont dam was selected as a real structure and optimized under different conditions against the Centro earthquake. The calculation parameters of the PSO algorithm showed the proper performance of this algorithm.

To check the random nature of the optimization algorithm, four independent executions were performed for the PSO method and their results were analyzed separately. The results showed that with the number of 10,000 analyzes, the volume of concrete used was equal to 346,000 m3.

The objective of distribution and delivery of water canal scheduling in irrigation canal networks is timely and adequate delivery of water with minimum operational stages of the head gate of supply canal in the presence of canal capacity and irrigation rotation period constraints. Poor performance of irrigation canals and its effect on decreasing of Agricultural water productivity requires attention for their improvement. Traditional approach for water delivery planning is based on personal experiences, which is not necessarily Satisfactory. The use of analytical and optimization methods could resolve some of these difficulties. Classical optimization methods are facing some limitations such as: being trapped in local optimum points, and difficulties in handling different variables. To overcome some of these limitations, new techniques which can solve complex problems could be used. This study has used the meta-optimization algorithm of PSO optimization to distribute and deliver optimal water in the irrigation network of Nesa Dam in Bam city located in Kerman province. In this research the delivery and distribution program in distribution channel branches are provided so that the various objects such as decreasing in distributor channel capacity and decreasing in time needed for complete the irrigation program optimize as a single and two objectives. In this program, first branch numbers, upper and lower limit of delivery discharge to per branch and branch coverage, gross irrigation requirement, irrigation interval and block numbers as input are defined to the model. By running the model, the best intermittent of branches in per block, minimum of distributer channel capacity and minimum irrigation duration in optimum conditions define to the model as outputs. According to the results presented the maximum flow rate is 2469 lit/s and the maximum irrigation time is 356 hours, which is reduced by 531 lit/s and 4 hours of irrigation time.

Canal systems of distributing water have different design capacities, command areas and the lengths requiring different duration of operation. Irrigation scheduling becomes a complex process under these conditions, especially for rotational water distribution. Planning for optimal operation in irrigation network is an efficient solution in water saving, increasing in cultivation and improving agriculture performance. Delivery and distribution in irrigation network are multi-objective and have multi variation optimization and multi limitation problems. By the optimal delivery schedule, minimum canal capacity and minimum delivery time is derived. Minimization of canal capacity reduces the construction costs and minimization of delivery time, limits water losses due to seepage, evaporation and water stealing. Solving such problem needs powerful optimization methods. Evolutionary algorithms are the major group of search algorithms for finding a suitable answer, using current evolutionary rules in nature. In this research ICA and PSO algorithms are applied. To achieve the research goals, the considered algorithms are applied in MATLAB 2014 software. The results of this research are also compared with GA method in study of Monem et al, (2007) and Ants Colony System Algorithm in study of Emadi and Kakouie (2012).
Imperialism is the policy of extending the power and rule of a government beyond its own boundaries. ICA is a novel global search strategy that uses imperialism and imperialistic competition process as a source of inspiration. This algorithm is based on that in real world countries try to extend their power over other countries to use their resources and bolster their own government. In fact, imperialist countries attempt to dominate other countries and turning them to their colonies. Also, imperialists compete strongly with each other for taking possession of other countries. During this competition stronger empires get more power and the weakest one ultimately collapses. This general policy of imperialist competition is used as the basis of the ICA. Particle swarm optimization (PSO) is a population-based stochastic approach for solving continuous and discrete optimization problems. In this research, the delivery and distribution program in distribution channel branches with ICA and PSO algorithm were provided so that the various objects(such as decreasing in distributor channel capacity and decreasing in time) needed for complete the irrigation program optimize, as a single and two objectives. In this program, defined inputs to the model were: first branch numbers, the upper and lower limit of delivery discharge to each branch and the branch coverage, gross irrigation requirement, irrigation interval and block numbers. By running the model, the best intermittent of branches in per block, minimum of distributer channel capacity and minimum irrigation duration in the optimum conditions define to the model, as outputs. In order to comparison ICA and PSO result with ACS and GA, the model is applied on BP14 canal in Fomanat Irrigation Network in west Gilan Province, Iran. This canal has a length of 6852 meters with a trapezoidal cross section and capacity of 2.5 m3/s and land cover area of 2100 ha. Irrigation water requirement is 50 mm.
The optimal delivery schedule canal for single objective (canal capacity minimization) is derived as Follow: canal capacity: 595 lit/s, time delivery: 343hr and the number of upstream gate regulations: 12 in PSO algorithm; and canal capacity: 546 lit/s, time delivery: 357hr and the number of upstream gate regulations: 11 in ICA. The optimal delivery schedule canal for two objective including minimization canal capacity and delivery time is derived as follow: canal capacity: 645lit/s, time delivery: 300hr and the number of upstream gate regulations: 12 in PSO algorithm; and canal capacity: 630lit/s, time delivery: 292hr and the number of upstream gate regulations: 12 in ICA.
ICA had the better Optimal Water Scheduling comparing with GA, ACS and PSO. Comparison of results obtained from ICA with other method's result showed that ICA in both single and two objective state determined less canal capacity. Also, in ICA the number of upstream gate regulations was one less than other approaches. It was found that PSO and ACS had the best performance in the next level. So, ACS determined less canal capacity and PSO determined more irrigation schedule complement duration. Therefore, ICA algorithm is a powerful method that able to escape from the local peaks and reach to global optima. ICA algorithms more likely reach Global optimum positions than PSO and GA. This algorithm, due to wide space search, could create a better collocation for branches of canal in irrigation network which leads to the better optimization.

Planning for optimal operation in irrigation network is an efficient solution in water saving, increasing in cultivation and improving agriculture performance. Problem of water delivery scheduling in irrigation network is a multi-objective, multi-variable, and multi-constraint problem. Solving such problem needs powerful optimization methods. Evolutionary algorithms are major group of search algorithms that are used for finding a suitable answer using current evolutionary rules in nature. In this research the provided delivery and distribution program in irrigation network is optimized using Imperial Comparative Algorithms (ICA) and PSO and are compared with genetic algorithm results. In this research the delivery and distribution program in distribution channel branches are provided so that the various objects such as decreasing in distributor channel capacity and decreasing in time needed for complete the irrigation program optimize as a single and two objectives. In this program, first branch numbers, upper and lower limit of delivery discharge to per branch and branch coverage, gross irrigation requirement, irrigation interval and block numbers as input are defined to the model. By running the model, the best intermittent of branches in per block, minimum of distributer channel capacity and minimum irrigation duration in optimum conditions define to the model as outputs. To control and considering of developed program, the optimum delivery program in downstream irrigation network of Jiroft diversion dam using this model is provided. The results show that in two-objective optimum status for irrigation supply of the network, Imperialist Competitive Algorithm performance is better than the other two algorithms so that the maximum discharge is 1381 cubic meter per second and maximum irrigation duration is 236 hours.

Determination of water distribution uniformity of coefficient of sprinkler irrigation based on data of a single sprinkler is time consuming due to overlapsprinkling by neighboring sprinklers and also different pressure heads, riser heads, sprinkler gaps on laterals and the distance between laterals. The best combination of the mentioned parameters to achieve maximum water distribution uniformity of coefficient, is still unknown question for applicators. In this research, water distribution uniformity of coefficient of ZB model sprinkler (made in Iran) were measured at Hashemabad cotton research station of Gorgan under 3 different pressure heads (2.5, 3 and 3.5 atm), two riser heads (60 and 100 cm) and seven sprinkler (Sl×Sm including 9×12, 9×15, 12×12, 15×12, 12×18, 15×15, 15×18m) arrangements. Two different algorithms namely PSO and GA were developed in MATLAB to determine water distribution uniformity of coefficient parameters with respect to mentioned parameters. based on statistical parameters such as coefficient of determination (R2), root mean square error (RMSE) and also standard deviation (SD) the performance of these algorithms were investigated. The results show that both of them can accurately predict water distribution uniformity of coefficient and therefore they can be used in determination of water distribution uniformity of coefficient.

Soil erosion is a major environmental threat to sustainability and agricultural productivity leading to reduction of soil fertility, loss of nutrients, and declines of crop yields in arable lands. In this study, the suitability of SWAT model for simulating the daily runoff and sediment in the Bazoft watershed (one of the main sub-basins of the Karun river basin in central Iran) was investigated. For this purpose, three different climate data sets including climate data obtained from the Bazoft watershed (Bazoft watershed data), the climate research unit data (CRU data), and a combination of the CRU and Bazoft watershed data (Bazoft-CRU data) were first investigated. Then, the potential use of PSO algorithm for calibrating and validating of the model was evaluated. The constructed SWAT model using a combination of the CRU data and the Bazoft climate data had a better runoff simulation performance in the study area. The obtained Nash-Sutcliff (NS) coefficient for the constructed SWAT model using the Bazoft-CRU data was 0.38 while it was 0.31 and 0.21 for the proposed SWAT model using the Bazoft watershed data and the CRU data, respectively. The obtained r-factor and p-factor values in the model calibration period using the PSO algorithm were 1.25 and 0.81, respectively. The R2 and NS coefficients for the validation period were also 0.64 and 0.60, respectively. Furthermore, in the sediment simulation step, the r-factor and p-factor values for the calibration period were 0.69 and 0.85, respectively, and they were 0.63 and 0.80 for the validation period, respectively.

Optimal design of water distribution system is a real problem and its aim is finding the best solution for transferring of water from the reservoir to consumers;so that, all their needs to be supplied with the lowest cost. Many researchers proposed various researches such as linear, nonlinear programming, global optimization,and evolutionary algorithms for minimizing the cost of these systems. The Particle Swarm Optimization (PSO) algorithm, which is an evolutionary algorithm originally introduced for solving optimization problem with continuous variables, but later it was developed for problems with discretevariables. Inthis research, for minimizing the cost of waterdistribution system design after modifications in continuousPSO algorithm, the program was linked to the hydraulic simulator (EPANET 2.0 in Matlab environment). Finally, for improving the algorithm performance, and obtaining the best values of PSO in optimization and water distribution system design, we were used three benchmark networks from previous studies that are different in layout and also the number of pipes, which they called Two-loop, Hanoi, and Kadu networks. From Two-loop network optimization, the least-cost network was estimated from 3100 number of objective function evaluation, which caused to improve the results, when compared to the pervious algorithms. In Hanoi network, the minimum cost was obtained $6.097×106 with =10.667 value, which was less than the results of other networks. Additionally, in Kadu network the cost was obtained 130,666,043 Indian rupees, which was less when compare to the other methods. The results showed a good performance of this algorithm when compared to the other algorithms.