One of the most important processes in the earlystages of construction projects is to estimate the costinvolved. This process involves a wide range of uncertainties, which make it a challenging task. Because ofunknown issues, using the experience of the experts orlooking for similar cases are the conventional methods todeal with cost estimation. The current study presents datadrivenmethods for cost estimation based on the applicationof artificial neural network (ANN) and regression models. The learning ALGORITHMS of the ANN are the Levenberg– Marquardt and the Bayesian regulated. Moreover, regressionmodels are hybridized with a genetic algorithm toobtain better estimates of the coefficients. The methods areapplied in a real case, where the input parameters of themodels are assigned based on the key issues involved in aspherical tank construction. The results reveal that while ahigh correlation between the estimated cost and the realcost exists; both ANNs could perform better than thehybridized regression models. In addition, the ANN withthe Levenberg– Marquardt learning algorithm (LMNN)obtains a better estimation than the ANN with the Bayesian-regulated learning algorithm (BRNN). The correlationbetween real data and estimated values is over 90%, whilethe mean square error is achieved around 0. 4. The proposedLMNN model can be effective to reduce uncertainty andcomplexity in the early stages of the construction project.

A number of ALGORITHMS based on the evolutionary processing have been proposed for communication networks backbone such as Genetic Algorithm ((GA)). However, there has been little work on the SWARM optimization ALGORITHMS such as Particle Swarm Optimization (PSO) for backbone topology design. In this paper, the performance of PSO on (GA) is discussed and a new algorithm as PSO(GA) is proposed for the network topology design. The simulations for specific examples show that the performance of the new algorithm is better than other common methods.

Today, one of the important issues in the industry is the failure of parts due to the presence of holes or cracks. Among the numerical calculation tools, the classical and extended finite element method is known as the most useful numerical tools in solving engineering science problems. Identifying and investi(GA)ting the types of cracks, flaws and cavities in structures is one of the most challenging issues in the field of engineering. In this article, the crack detection of two-dimensional (2D) structures using the extended finite element method (XFEM) along with genetic algorithm((GA)) and grey wolf optimization method (GWO) to detect the existing crack and flaws by minimizing an error function which is also called as objective function that the evaluation of it, is based on difference between sensor measurements and suggested structure responses in each try of the algorithm.  Damage detecting in 2D domains, as a non-destructive evaluation problem, is investi(GA)ted using the extended finite element method along with the optimization method of genetic algorithm and grey wolf. The extended finite element method has been used to model the structure containing cracks and holes in the abaqus program, and genetic optimization and grey wolf method have been used to determine the location of the damage in which the codes were in matlab program. The extended finite element method is a powerful tool for the analysis of structures containing cracks without remeshing and is therefore suitable for an iterative process in structural analysis. Also, in these problems, due to the wide range of parameters, it is not logical and rational to use mathematical methods. For this reason, meta heuristic methods have been developed, and grey wolf optimization methods and genetic algorithm are among these common non-gradient methods that are suitable for solving the inverse problem. This problem is set so that the optimizer algorithm finds the existing crack coordinates or holes coordinates by minimizing an objective function based on the values measured by the sensors installed on the structure. Among the limitations of the classical finite element method in the investi(GA)tion of various problems in the field of fault and crack detection, we can point out the dependence of the crack or cavity on the finite element mesh, re-meshing and in other special cases the use of singular elements, which are completely removed by using The extended finite element. In this research, in order to identify the damage, the genetic optimization algorithm and the gray wolf have been used. These ALGORITHMS are designed in such a way to determine the characteristics of the damage by minimizing an error function. The defined error function is defined as the difference between the response obtained from the algorithm analysis and the response recorded in the main structure modeled in ABAQUS software, at the location of the sensors. Finally, three reference numerical examples have been solved to evaluate the capability and accuracy of the proposed method, and the result of the results shows a reduction in the cost of solving and an increase in the accuracy of the results.

The powerfulness and effectiveness of the optimization methods are motivations of the researchers to use them in complex engineering problems. In this paper, the performance of the three optimization ALGORITHMS based on swarm intelligence (IPO, PSO) and evolutionary technique ((GA)) for calculation the channel's widths of the transistors were evaluated compared with each other’s. The fitness functions are defined in order to the better integration and to improve the power consumption and delay of Level Shifter circuit (LS) with changing the voltage level of 0.4 to 3 volts using 0.35-um CMOS technology. Simulation results for the sample circuit show that it reach a power consumption of 0.222pW and a delay value of 9.113ns with PSO algorithm, a power consumption of 0.39 nW and delay value of 3.741 ns with IPO algorithm, and values of 0.235 nW and 3.711 ns whit (GA) algorithm. In addition to a dramatic improvement in power and delay, minimum of channel's widths also were obtained. All implementations of paper were performed in MATLAB and HSPICE.

Urban transportation is one of the most important issues of urban life especially in big cities. Urban development, and subsequently the increase of routes and communications, make the role of transportation science more pronounced. The shortest path problem in a network is one of the most basic network analysis issues. In fact, finding answers to this question is necessity for higher level analysis. In general, shortest path solution methods using optimization ALGORITHMS are divided into two categories: exact and approximate ALGORITHMS. In exact ALGORITHMS, achieving the optimal solution requires time, and consequently more cost. On the opposite side, there are some approximate ALGORITHMS that work in a short period of time. Meta-heuristic ALGORITHMS are among approximate ALGORITHMS that are capable of finding optimal or near-optimal solutions in a reasonable period of time. The method used in this study is to solve the shortest path problem with the combination of Genetic meta-heuristic ((GA)) and Tabu Search (TS) ALGORITHMS. (GA) is inspired by genetic science and Darwin's theory of evolution; it is based on survival of the highest or natural selection. A common use of genetic ALGORITHMS is to be used as an optimization function. In (GA), the genetic evolution of living things of life is simulated. Inspired by the evolutionary process of nature, these ALGORITHMS solve problems. (GA) forms a set of population (solutions), then it achieves an optimal set by acting some possess on the correct set. To solve a problem by genetic ALGORITHMS, it is necessary the problem is converted to the specific form required by (GA). On the other hand, TS algorithm is not population-based. It obtains an answer, then it tries to direct the answer to the optimal solution by applying a series of operators. This algorithm is highly similar to the Simulated Annealing algorithm. In this paper, for solving the shortest path problem, a series of geometric pre-processing on the network is done to generate a search area around the source and destination nodes. In the proposed algorithm, the cost function is defined as a complex number, which the real part shows the sum of the weight of the real edges, and the imaginary part denotes the number of virtual edges. The innovation of this research is about applying Tabu Search algorithm in mutations process of genetic algorithm. The proposed method overcomes the inappropriate response of the pure genetic algorithm in terms of the final weight of the path especially the large networks. In order to evaluate the efficiency of the proposed algorithm, the algorithm was implemented on a real directional network which is part of Tehran city road networks including 739 nodes and 1160 edges. The results show that in the proposed algorithm, the length of the path is as close as possible to the solution obtained from the definitive Dijkstra’ s algorithm. This algorithm predicts approximately the final path length of 5% more than Dijkstra’ s algorithm. But in terms of running speed, it is 5. 12 times faster than the Dijkstra’ s algorithm. In comparison with the pure genetic algorithm, the proposed algorithm is 9% shorter in average in terms of path length. And about the running time, the speed of the proposed algorithm is approximately equal to the pure genetic algorithm. Re(GA)rding to repeatability, the proposed algorithm also shows 25. 36% of repeatability.

Background and Objectives: Optimal utilization of water resources systems and appropriate formulation rules and policies for the exploitation of reservoirs have been considered by water resource experts in recent years and extensive research has been carried out. Although much progress has been made in terms of problem-solving strategies and computational tools, the problem of optimizing the operation of a multi-reservoir systems due to the effect of upstream storage capacities on downstream inputs, is so complicated. Routine optimization methods due to high constraints, discontinuous space and non-linear nature of water resource management issues are not a good tool for solving such problems. For this reason, the metaheuristic optimization ALGORITHMS have been considered by researchers. Materials and Methods: In this research, the performance of the (GA) and PSO in solving the problem of optimizing the operation of a multi-reservoir system including Bostan and Golestan dams located in Gor(GA)n-Rood watershed has been studied and compared. The survey of the input to the two dam reservoirs in the year 2014-2015 shows that due to the climate change, the annual input to the Bostan and Golestan dams has decreased by 17% and 60%, respectively. Genetic algorithm is a parallel and guided search based on the theory of evolution. The operators of the (GA) algorithm include selection, crossover and mutation that are used to make up the next generation, respectively. In PSO optimization algorithm, based on the birds and fishes movements, a number of particles are propa(GA)ted in the search space and the value of the objective function is calculated in proportion to the position of each particle. Then the new particle position is calculated using the combination of current particle locations and the best place previously used. Results: The best answer of the PSO algorithm during the 10 runs is 909. 95 and the worst is equal to 930. 53, while the best answer of the (GA) algorithm during the 10 runs is 931. 17 and the worst was 957. 32. The comparison of the mean of the answers also show that the PSO algorithm has a 3% advantage over (GA). Conclusion: The PSO algorithm has a better performance than (GA), so that the PSO algorithm with a reliability of 49. 38% has a better performance than the (GA) algorithm with a reliability of 48. 44%.

With respect to necessity of the optimum use of water resources and existence of many various optimization methods, in this study 3 kinds of heuristic ALGORITHMS have been used including Particle Swarm Optimization, Genetic Algorithm and Simulated Annealing to optimize the operation of Shaharchai dam reservoir as an application. The optimization was carried out considering the probability of inflow for a period of 5 years. In order to obtain the best operation of reservoir, monthly release was defined as a second order polynomial according to storage volume and inflow, and different parameters of these ALGORITHMS have been adjusted to minimize the objective function in which supplying the required demand of downstream was defined as the target. The best state of each algorithm is selected through 10 times running of programs (due to intrinsic random behavior of ALGORITHMS) and the results comparison leads to realization of which method can perform the best. According to the results, Particle Swarm Optimization method operates more effectively and produces the best results in solving reservoir operation problems. So as an application, control curves of release and storage volume have been extracted for Shaharchai dam reservoir using this method.

A total of 1099 data points consisting of alcohol-alcohol, alcohol-alkane, alkane-alkane, alcohol-amine and acid-acid binary solutions were collected from scientific literature to develop an appropriate artificial neural network (ANN) model. Temperature, molecular weight of the pure components, mole fraction of one component and the structural groups of the components were used as input parameters of the network while the refractive index was selected as its output. The ANN was optimized once by genetic algorithm ((GA)) and once a(GA)in by particle swarm optimization algorithm (PSO) in order to predict the refractive index of binary solutions. The optimal topology of the ANN-(GA) consisted of 13 neurons in the hidden layer and the optimal topology of the ANN-PSO consisted of 16 neurons in the hidden layer. The results revealed that the ANN optimized by PSO had a better accuracy (MSE=0. 003441 for test data) compared to the ANN optimized with (GA) (MSE=0. 005117 for test data).