evolutionary algorithms

Considering the growing importance of energy storage technologies in large-scale construction to improve the performance of power grids, this paper examines the investment in these technologies from an economic perspective. The main article focuses on the use of large battery technologies such as redox vanadium batteries, polysulfide-bromine batteries, and zinc-bromine batteries for power grids due to their special characteristics such as long life, good efficiency, and affordable costs. Smart are very suitable. The proposed model of this research is designed using golf optimization algorithm, which is considered as the main innovation of this research. This algorithm is specifically used to optimize performance and investment in large-area energy storage systems. Using this model, the effects of using different types of technologies are analyzed. The results of the simulations show that considering different battery technologies and the operation of the golf optimization algorithm can have significant economic effects in optimizing the performance of energy storage systems.

The 3D integrated circuit is emerged as a promising solution to integrate very large-scale circuits on electronics chips. In such chips, several layers of silicon substrates are stacked which are separated by insulator interfaces. Interconnection between two layers is realized using Through Silicon Via (TSV). Fabrication of TSVs is challenging due to their large size and complex process. Consequently, the number of TSVs should be minimized in the circuit’s implementation. The 3D implementation consists of three main steps: Partitioning, Placement, and Routing. In this paper, the first two steps are accomplished using the Simulated Annealing-based optimization approach wherein minimization of the number of TSVs and total wire length are considered the main objectives. In this paper, an improved version of the pathfinder method has been developed which would efficiently generate the necessary interconnections among circuit modules. The results of simulations on MCNC benchmark circuits show that the proposed method outperforms the previous state-of-the-art methods in all aspects. In comparison with FSA, the number of TSVs is reduced by 6.15%, and the algorithm’s runtime is decreased by 27.79%. Moreover, in comparison with the hMETIS method, the number of TSVs is reduced by 9.78%, and the algorithm’s runtime is decreased by 31.73% .

The safety analysis of software systems, especially safety-critical ones, should be performed exactly because even a minor failure in these systems may result in disaster consequences. Also, such analysis must be done before implementation, i.e. the design step and in the model level. Model checking is an exact and mathematical-based way that gets a model of a system and analyzes it through exploring all reachable states of the model. Due to the complexity of some systems and their models, this way may face the state space explosion problem, i.e. it cannot explore all available states. A solution to solve this problem in these systems is that model checking tries to refute them, instead of verifying them, by finding errors such as deadlock (if available).Although, a heuristic has been previously proposed to find a deadlock in the model's state space and it has been applied in several simple heuristic search and evolutionary algorithms, its detection speed has been low. In this paper, we propose a novel heuristic to detect a deadlock in the model's state space, and test and compare its detection speed by applying it in several simple heuristic search algorithms such as iterative deepening A*, beam search, and evolutionary algorithms such as genetic, particle swarm optimization, and Bayesian optimization. Comparison results confirm that the new heuristic can detect a deadlock in less time than the previous heuristic.

One of the most challenging issues with web services is the composition of them, which is presented as a graph to show the interaction between services. Each node in this graph is called an abstract web service which their function is specified but the quality features are unclear. For each abstract service, there is a set of candidate services with the same function but different qualitative features. Replacing a candidate web service for each abstract service so that an optimal combination is achieved is an NP-hard problem that cannot be solved in polynomials, hence to solve it using heuristic algorithms. Several methods have been proposed for the web services optimal composition, but most of these methods don't support the probability structure. Only one method supports a probability structure that is not scalable for large graphs, constraint based, and analyzes each path of the graph separately. This paper presents an integrated scalable multi-objective approach for analyzing graph to not only covering two new patterns of nested loops and parallel loops, but also improving performance with representing a method for simplifying web-service composition. In this method, evolutionary algorithms are used for optimal web services selection and scalability. The two selected evolutionary algorithms are NSGAII and SPEAII. In the proposed method, first in conditional graphs, each path is repeated according to its probability, and then the NSGAII algorithm is used to determine the best path in the graph and find better solutions.  At the end of the article, the results of 8 methods are presented, which compared to the best of them, the proposed method has improved 30% in the reliability parameter and 121 milliseconds in response time.

In this paper, to solve the multi-objective problem of distribution feeder reconfiguration (DFR) in the presence of distributed generation (DG), the hybrid particle swarm optimization and gravitational search algorithm (PSOGSA) has been proposed, which is a combination of particle optimization (PSO) and gravitational (GSA) optimization algorithm. In this field, the power losses and operating costs are the two most used objective functions in the literature. In addition to the mentioned objective functions, this paper also considers the optimal generation capacity of DG resources and energy not supplied (ENS), which is one of the basic reliability indexes of distribution networks. In this paper, the values of different objective functions are normalized by the fuzzy method, and also the Fuzzy decision-maker is used to determine the most optimal solution among the Pareto-optimal solutions. The proposed algorithm is implemented on IEEE 70-bus and 119-bus test systems. The simulation results show the efficiency of the proposed PSOGSA in improving the considered objective functions. The proposed method, by establishing a suitable fit between different objective functions has introduced a more efficient structure with lower losses and operating costs, as well as greater reliability, compared to other optimization algorithms.

Quantum computing is a new method of information processing based on the concepts of quantum mechanics which leads to strange and powerful events in the field of quantum. Each unitary matrix represents a quantum gate. Synthesis of quantum circuits refers to the process of converting a quantum gate into a series of basic gates and is divided into two general categories, namely decomposition and composition-based. In the first category, quantum circuits are synthesized using matrix decomposition methods. This study deals with the second category, which uses evolutionary algorithms and especially genetic algorithms to multiply matrix gates to achieve the final characteristic of a quantum circuit.

Due to the growing desire of people to join and use social networks, communication and shar ing data in these networks has been considered by various sciences such as political science, psychology, sociology, economics, etc. Hence, researchers have begun to distinguish and extract relationships between individuals from the data contained in these networks, to create more accurate communities. However, there is still no effective method to identify and extract communities based on social media data. In this article, a method has been proposed for social network accurate clustering by using Imperial Competitive Evolutionary Algorithm (ICEA) and selecting the initial population based on the density-based clustering criterion. The proposed method has improved the result of modularity about 21.45% in average, compared to rival basic ICEA and extracted more densed communities.

Maneuvering air targets Tracking has many applications in defensive and non-defensive areas. Target tracking requires the estimation of position, velocity, and acceleration, simultaneously. Common approaches for air targets tracking measures the distance to target and target heading angle which is a nonlinear function of system states. Since these measurements are noisy, using estimating and filtering methods for assessing the speed and acceleration of target is essential. Although, extended Kalman filter works well with nonlinear systems and Gaussian noises; in practice we encounter with non-Gaussian noises such as Glint which particle filters show better performance in them. In addition, due to the high computational load of particle filters, implementing and applying them is impossible. In this paper, the evolutionary algorithm of particle swarm optimization has been used in the sampling step to reduce computational load and improve the real-time performance of particle filter in solving air target tracking problems. The proposed method is simulated and evaluated in a scenario involving all possible motions of the target with high maneuvering. In addition, performance of an evolutionary particle filter with some particle filters and extended Kalman filter is compared. The simulation results indicate that the evolutionary particle filter has the capability of real time air targets tracking with maneuvering in comparison with the particle filter and extended Kalman filter while having high precision.

Wireless sensor networks have a large number of limited-energy sensor nodes dispersed in a finite area. Most node energies are used to send data to the central station. Due to the energy constraints in this type of grid, increasing life expectancy has always been a concern with decreasing energy consumption. The aim of this study is to provide surface clustering based on genetic algorithm in order to increase the life span of these networks. In proposed surface clustering, the geographic area is divided into three levels according to the radio range and the clustering of the nodes of each level is done individually. The cluster heads use more energy than other nodes to send information, so the proposed algorithm aims to reduce the number of cluster heads in order to increase the network lifetime. Finally, by changing the clusters in each routing round, there is a greater energy balance between the nodes. The results from the experiments indicate the superiority of the proposed algorithm in transmitting messages and network lifetimes over other similar protocols.

One of the most important problems for power system operation is unit commitment (UC), for which different constraints should be satisfied. UC is a nonlinear and large-scale problem; thus, using the evolutionary algorithms has been considered for solving the problem. In this paper, the solution of the UC problem was investigated using Modified Imperialistic Competition Algorithm (MICA).  Simulations were performed for a 10, 60 and 100-unit IEEE test system to produce the demand energy during a period of 24-hour. The obtained results were compared with those of some pervious algorithms such as GA, ICGA, PSO and their modified versions, and Cuckoo searching. The comparisons demonstrated the economic advantage of the presented method.

The growing use of energy in the world necessitates the development of power networks. However, developing new transmission lines requires a great deal of time and cost, so it will be very cost-effective to use the same lines with higher transmission capacities, if possible. In this regard, in recent years, by introducing of FACTS to power networks, their use in industrialized countries has become commonplace to increase the capacity of transmission lines. In this paper, the optimal adjustment of reactive power sources in the power network with FACTS series and parallel devices (TCSC, SVC) in order to coordinate them with each other and using fuzzy logic based on evolutionary algorithms such as particle swarm to reduce power losses Active, operating system costs including the cost of FACTS devices and congestion in transmission networks. Finally, this will be proven by simulating the IEEE 30-Bus test network and placing FACTS devices on it.

Prediction of chaotic time series based on the phase space reconstruction theory has been applied in many research fields. In this paper, we propose an improved adaptive neuro-fuzzy inference system (ANFIS) with self-feedback and imperialist competitive learning algorithm for the application of chaotic time series prediction. Since the ANFIS is based on a feed-forward network structure, it is limited to static problems and cannot effectively cope with dynamic properties such as the time series. To surmount this trouble, we suggested an improved version of ANFIS by introducing self-feedback connections from previous outputs that model the temporal dependence. Also we suggested a new hybrid learning algorithm based on imperialist competitive algorithm (ICA) and least square estimation (LSE) to train this new ANFIS structure. This hybrid learning algorithm is free of derivation and solves the trouble of falling in local optimum in the gradient based algorithm for training the antecedent part. The proposed approach is used to model and predict the six benchmarks of chaotic time series. Analysis of the prediction results and comparisons with recent and old studies demonstrates the promising performance of the proposed approach for modeling and prediction of nonlinear and chaotic time series.

Link prediction in social networks is one of the most important activities in analysis of such networks. The importance of link prediction in social networks is due to its dynamic nature. While members and their relationships (links) in such networks are continuously increasing, links may be missed due to various reasons. By predicting such links, the possibility of extension, completion, and information retrieval of these networks can be provided. For predicting and detecting these links, some information is needed, which they could be extracted from the network graph structure and used as measures for prediction. Having introduced two new measures, which lead to an effective performance in predictions and recommendations, we propose a new method. The link prediction is achieved through combining classifiers and evolutionary methods of GA (Genetic Algorithm) and ICA (Imperialist Competitive Algorithm). To show practical considerations, two real datasets, Facebook and Epinions, are employed. We show that the proposed method can increase the performance and precision of prediction.

One of the most effective techniques for energy management in wireless sensor networks is clustering. How to establish clusters and the method used to choose cluster heads are the most important factors in creating optimal clusters. In fact, appropriate cluster head selection results in longer network lifetime and more data delivery to the sink. In this paper, after exploring clustering algorithms based on computational intelligence techniques and evaluating the strengths and weaknesses of each, we provide two novel clustering schemes termed as Energy Aware Genetic Clustering Algorithm (EAGCA) and EAGCA*. In these centralized algorithms, we try to achieve an optimal clustering by using global information such as number of neighbor nodes, the energy of neighbor nodes, local distance and local traffic load on each node. Compared to some well-known clustering algorithms such as LEACH and EAERP, simulation results demonstrate that EAGCA and EAGCA* are able to create more appropriate clusters in terms of energy consumption, life time and total number of received data to the sink.

In this paper، a new approach for selective harmonic elimination (SHE) in a cascaded multilevel inverter is proposed. The switching angles are determined with the assumption of varying input DC sources and at this condition the fundamental component is remained adjusted and undesired harmonic components are eliminated. The on-line switching angles determination is done by an Artificial Neural Network (ANN). The training data set for ANN are produced by application of heuristic algorithms to solve the SHE problem. In this paper، various algorithms such as harmony search algorithm، imperialist competitive algorithm and particle swarm optimization are applied to obtain the switching angles and by comparing the results of these methods، the better one is used for ANN training.