University of KashanSoft Computing Journal2322-37078220210523An approach to Improve Particle Swarm Optimization Algorithm Using CUDAAn approach to Improve Particle Swarm Optimization Algorithm Using CUDA22111144210.22052/8.2.2FAMohammad PouyaAkbarpourKeyhanKhamforooshVafaMaihamiJournal Article20210523The time consumption in solving computationally heavy problems has always been a concern for computer programmers. Due to simplicity of its implementation, the PSO (Particle Swarm Optimization) is a suitable meta-heuristic algorithm for solving computationally heavy problems. However, despite the simplicity, the algorithm is inefficient for solving real computationally heavy problems but the presence of local interactions between particles has made this algorithm suitable for parallelization. On the other hand, by the invention of GPU (Graphical Processor Unit) and introducing the CUDA architecture as a GPU in the NVIDIA graphical processor, fundamental changes has been made in solving this type of problems. Despite all the research done in the field of implementing the algorithms through GPUs, some aspects of parallelization have not been addressed for suitable speedup and efficiency on NVIDIA GPUs. By considering the Geforce GT 525M, which is a relatively weak GPU, this paper tries to achieve the maximum speedup of the algorithm by implementing on this GPU. This experience led to reaching the acceptable efficiency on other GPUs. To reach the achievement, the multi-kernel model was used. The results show the speedup of 15.98 in solving the Rastrigin function.The time consumption in solving computationally heavy problems has always been a concern for computer programmers. Due to simplicity of its implementation, the PSO (Particle Swarm Optimization) is a suitable meta-heuristic algorithm for solving computationally heavy problems. However, despite the simplicity, the algorithm is inefficient for solving real computationally heavy problems but the presence of local interactions between particles has made this algorithm suitable for parallelization. On the other hand, by the invention of GPU (Graphical Processor Unit) and introducing the CUDA architecture as a GPU in the NVIDIA graphical processor, fundamental changes has been made in solving this type of problems. Despite all the research done in the field of implementing the algorithms through GPUs, some aspects of parallelization have not been addressed for suitable speedup and efficiency on NVIDIA GPUs. By considering the Geforce GT 525M, which is a relatively weak GPU, this paper tries to achieve the maximum speedup of the algorithm by implementing on this GPU. This experience led to reaching the acceptable efficiency on other GPUs. To reach the achievement, the multi-kernel model was used. The results show the speedup of 15.98 in solving the Rastrigin function.https://scj.kashanu.ac.ir/article_111442_d31878ac8034256c3ff3ea74afcbd25f.pdfUniversity of KashanSoft Computing Journal2322-37078220210523Improvement of the Analytical Queries Response Time in Real-Time Data Warehouse using Materialized Views ConcatenationImprovement of the Analytical Queries Response Time in Real-Time Data Warehouse using Materialized Views Concatenation223711144310.22052/8.2.22FASeyed MajidShafaeiBabakVaziriSeyed MoostafaShafaeiJournal Article20210523A real-time data warehouse is a collection of recent and hierarchical data that is used for managers’ decision-making by creating online analytical queries. The volume of data collected from data sources and entered into the real-time data warehouse is constantly increasing. Moreover, as the volume of input data to the real time data warehouse increases, the interference between online loading operations and online analytical processing increases. These two stated challenges have become the most important issues regarding real time data warehouse. In this article, a method is presented to improve the analytical queries response time in the real time data warehouse architecture using materialized views concatenation. This process takes place by: (1) storing the results of performed queries in each real time section, (2) transferring the results to the next section when transferring data to the section. Each real time section contains data of its previous section, which have been transferred in several stages. As a result, the calculated results of the queries are also transmitted by transferring data, and consequently, for achieving desired outcome, the previously calculated results can be combined without the need to run the queries again. The proposed method has reduced both analytical queries response time and data entry interference caused by the simultaneous and long-term execution of queries. This study faces two challenges: (1) applying the proposed method to a small amount of data in the real time section and (2) the changes in the proposed method for applying to big data.A real-time data warehouse is a collection of recent and hierarchical data that is used for managers’ decision-making by creating online analytical queries. The volume of data collected from data sources and entered into the real-time data warehouse is constantly increasing. Moreover, as the volume of input data to the real time data warehouse increases, the interference between online loading operations and online analytical processing increases. These two stated challenges have become the most important issues regarding real time data warehouse. In this article, a method is presented to improve the analytical queries response time in the real time data warehouse architecture using materialized views concatenation. This process takes place by: (1) storing the results of performed queries in each real time section, (2) transferring the results to the next section when transferring data to the section. Each real time section contains data of its previous section, which have been transferred in several stages. As a result, the calculated results of the queries are also transmitted by transferring data, and consequently, for achieving desired outcome, the previously calculated results can be combined without the need to run the queries again. The proposed method has reduced both analytical queries response time and data entry interference caused by the simultaneous and long-term execution of queries. This study faces two challenges: (1) applying the proposed method to a small amount of data in the real time section and (2) the changes in the proposed method for applying to big data.https://scj.kashanu.ac.ir/article_111443_7d577fbe898d27876c94598909b09ce3.pdfUniversity of KashanSoft Computing Journal2322-37078220210523Introducing a new meta-heuristic algorithm based on See-See Partridge Chicks Optimization to solve dynamic optimization problemsIntroducing a new meta-heuristic algorithm based on See-See Partridge Chicks Optimization to solve dynamic optimization problems386511144410.22052/8.2.38FAMajidMohammadpourBehroozMinaeiHamidParvinJournal Article20210523The SSPCO (See-See Particle Chicks Optimization) is a type of swarm intelligence algorithm derived from the behavior of See-See Partridge. Although efficiency of this algorithm has been proven for solving static optimization problems, it has not yet been tested to solve dynamic optimization problems. Due to the nature of NP-Hard dynamic problems, this algorithm alone is not able to solve such optimization problems. Therefore, to enable the algorithm to optimally track the variable in these problems, it is necessary to be provided solutions with this algorithm so that can increase the performance of this algorithm for dynamic environments. In this paper, two solutions for combining SSPCO are presented: (1) the multi-swarm method and (2) memory with Gaussian density estimation. The problem with most multi-swarm methods is that as the population increases uncontrollably, the speed and efficiency of the algorithm gradually decreases. The multi-swarm methods presented in this paper is adapted to the problem space, and whenever there is a need to increase the population, a population is created adaptively, and this reduces the problems of previous methods. One of the issues that is being addressed to solve uncertainty problems is prediction of near future using data of the near past. In this article, to preserve past data a new memory called Gaussian density estimation memory is used. This memory fixes standard memory defects and improves the performance of the proposed algorithm. To evaluate the efficiency of the proposed method, the well-known moving peak benchmark function, which simulates behavior of dynamic problems, is used. The proposed algorithm is compared with the 10 most popular dynamic optimization algorithms. According to the experimental results, the proposed method reduces offline error to a great extent compared to other methods and the error produced by the proposed method is very small.The SSPCO (See-See Particle Chicks Optimization) is a type of swarm intelligence algorithm derived from the behavior of See-See Partridge. Although efficiency of this algorithm has been proven for solving static optimization problems, it has not yet been tested to solve dynamic optimization problems. Due to the nature of NP-Hard dynamic problems, this algorithm alone is not able to solve such optimization problems. Therefore, to enable the algorithm to optimally track the variable in these problems, it is necessary to be provided solutions with this algorithm so that can increase the performance of this algorithm for dynamic environments. In this paper, two solutions for combining SSPCO are presented: (1) the multi-swarm method and (2) memory with Gaussian density estimation. The problem with most multi-swarm methods is that as the population increases uncontrollably, the speed and efficiency of the algorithm gradually decreases. The multi-swarm methods presented in this paper is adapted to the problem space, and whenever there is a need to increase the population, a population is created adaptively, and this reduces the problems of previous methods. One of the issues that is being addressed to solve uncertainty problems is prediction of near future using data of the near past. In this article, to preserve past data a new memory called Gaussian density estimation memory is used. This memory fixes standard memory defects and improves the performance of the proposed algorithm. To evaluate the efficiency of the proposed method, the well-known moving peak benchmark function, which simulates behavior of dynamic problems, is used. The proposed algorithm is compared with the 10 most popular dynamic optimization algorithms. According to the experimental results, the proposed method reduces offline error to a great extent compared to other methods and the error produced by the proposed method is very small.https://scj.kashanu.ac.ir/article_111444_274b7b3def9606c0fcf17fa113b3a550.pdfUniversity of KashanSoft Computing Journal2322-37078220210523Using the Particle Swarm Optimization Algorithm to Generate the Minimum Test Suite in Covering Array with Uniform StrengthUsing the Particle Swarm Optimization Algorithm to Generate the Minimum Test Suite in Covering Array with Uniform Strength667911144510.22052/8.2.66FASajadEsfandyariVahidRafeJournal Article20210523Up to now, several useful algorithms have been proposed to generate covering array, which is one of the branches of combinatorial testing. The main challenge in generating such arrays is generation of the arrays with a minimum number of test cases (for efficiency) at a proper time (for performance), for large systems. Covering array generation strategies are often divided into two general categories: computational and meta-heuristic. Computational strategies usually benefit high performance but have poor results in terms of efficiency. On the other hand, meta-heuristic strategies enjoy good efficiency but suffer low performance. Among the available strategies, the DPSO strategy generates the best results in terms of efficiency, but it does not benefit high performance; in contrast the GS strategy benefits good performance but has not good efficiency. Generally, there is no strategy that is good in terms of both efficiency and performance. In this paper, we try to produce an appropriate test suite of high efficiency and performance using PSO. A simple and effective minimizer function has also been used to increase the efficiency. The evaluation results show that the proposed solution has desirable outcomes in terms of efficiency and performance.Up to now, several useful algorithms have been proposed to generate covering array, which is one of the branches of combinatorial testing. The main challenge in generating such arrays is generation of the arrays with a minimum number of test cases (for efficiency) at a proper time (for performance), for large systems. Covering array generation strategies are often divided into two general categories: computational and meta-heuristic. Computational strategies usually benefit high performance but have poor results in terms of efficiency. On the other hand, meta-heuristic strategies enjoy good efficiency but suffer low performance. Among the available strategies, the DPSO strategy generates the best results in terms of efficiency, but it does not benefit high performance; in contrast the GS strategy benefits good performance but has not good efficiency. Generally, there is no strategy that is good in terms of both efficiency and performance. In this paper, we try to produce an appropriate test suite of high efficiency and performance using PSO. A simple and effective minimizer function has also been used to increase the efficiency. The evaluation results show that the proposed solution has desirable outcomes in terms of efficiency and performance.https://scj.kashanu.ac.ir/article_111445_9894c0e3997f85e559bbbd7552271d94.pdfUniversity of KashanSoft Computing Journal2322-37078220210523An Efficient Resource Allocation for Processing Healthcare Data in the Cloud Computing EnvironmentAn Efficient Resource Allocation for Processing Healthcare Data in the Cloud Computing Environment8010111144610.22052/8.2.80FAMostafaGhobaei-araniFatemehMahdi BabaeiJournal Article20210523Nowadays, processing large-media healthcare data in the cloud has become an effective way of satisfying the medical users' QoS (quality of service) demands. Providing healthcare for the community is a complex activity that relies heavily on information processing. Such processing can be very costly for organizations. However, processing healthcare data in cloud has become an effective solution to meet QoS demands of health users. In this paper, a fuzzy-based solution is presented for determining the optimal cloud using resource prediction technique. Besides, to make balance during the processing of tasks, based on fuzzy selector the virtual machine (VM) migration technique is used to migrate a VM from an overload server to an underload one. The proposed framework consists of two parts, local and global. To deliver the application to the global part, the local part must first be checked. If it is not suitable, the request will be delivered to the global part; indeed, the proposed framework works in a hierarchical manner. At first, a list of received requests is created and then using the proposed solution, the amount of available resources is estimated based on which the requested resources are allocated for processing. We used the Cloudsim toolkit to evaluate the proposed solution under various parameters and results have been compared with those of FAHP and ICA-K-Means algorithms. Compared to FAHP, the simulation results show that the proposed solution benefits from a 10% cost reduction and and a 12% reduction in cost compared to ICA-K-Means. Moreover, compared to FAHP and ICA-K-Means, the proposed method enjoys a reduction in number of rejected requests and an increase of 8% and 7% performance compared to the FAHP and ICA-K-Means, respectively.Nowadays, processing large-media healthcare data in the cloud has become an effective way of satisfying the medical users' QoS (quality of service) demands. Providing healthcare for the community is a complex activity that relies heavily on information processing. Such processing can be very costly for organizations. However, processing healthcare data in cloud has become an effective solution to meet QoS demands of health users. In this paper, a fuzzy-based solution is presented for determining the optimal cloud using resource prediction technique. Besides, to make balance during the processing of tasks, based on fuzzy selector the virtual machine (VM) migration technique is used to migrate a VM from an overload server to an underload one. The proposed framework consists of two parts, local and global. To deliver the application to the global part, the local part must first be checked. If it is not suitable, the request will be delivered to the global part; indeed, the proposed framework works in a hierarchical manner. At first, a list of received requests is created and then using the proposed solution, the amount of available resources is estimated based on which the requested resources are allocated for processing. We used the Cloudsim toolkit to evaluate the proposed solution under various parameters and results have been compared with those of FAHP and ICA-K-Means algorithms. Compared to FAHP, the simulation results show that the proposed solution benefits from a 10% cost reduction and and a 12% reduction in cost compared to ICA-K-Means. Moreover, compared to FAHP and ICA-K-Means, the proposed method enjoys a reduction in number of rejected requests and an increase of 8% and 7% performance compared to the FAHP and ICA-K-Means, respectively.https://scj.kashanu.ac.ir/article_111446_a561cb8231892cfa1fa0d59096e28d4c.pdfUniversity of KashanSoft Computing Journal2322-37078220210523An Energy Efficient Clustering Method using Bat Algorithm and Mobile Sink in Wireless Sensor NetworksAn Energy Efficient Clustering Method using Bat Algorithm and Mobile Sink in Wireless Sensor Networks10211511144710.22052/8.2.102FAShayestehTabatabaeiJournal Article20210523Wireless sensor networks (WSNs) consist of sensor nodes with limited energy. Energy efficiency is an important issue in WSNs as the sensor nodes are deployed in rugged and non-care areas and consume a lot of energy to send data to the central station or sink if they want to communicate directly with the sink. Recently, the IEEE 802.15.4 protocol is employed as a low-power, low-cost, and low rate communication standard for WSNs, which ensures real-time applications via Guaranteed Time Slot (GTS). Accordingly, in this paper, a new protocol energy efficient bat algorithm and a mobile sink are presented. It can choose the optimal route based on: (1) distance of a sensor node to sink, (2), bat loudness, and (3) energy level of its battery. Using the OPNET simulator version 11.5, the proposed method was simulated by bat and NODIC protocols, and IEEE 802.15.4 and the results were considered in terms of energy consumption, end to end delay, signal to noise ratio, the success probability of sending data to sink and rate of passing data. The results of the simulation showed the use of the above-mentioned parameters in the proposed method leads to the improvement of the network throughput against using the IEEE 802.15.4 protocol, Bat algorithm, and NODIC protocol.Wireless sensor networks (WSNs) consist of sensor nodes with limited energy. Energy efficiency is an important issue in WSNs as the sensor nodes are deployed in rugged and non-care areas and consume a lot of energy to send data to the central station or sink if they want to communicate directly with the sink. Recently, the IEEE 802.15.4 protocol is employed as a low-power, low-cost, and low rate communication standard for WSNs, which ensures real-time applications via Guaranteed Time Slot (GTS). Accordingly, in this paper, a new protocol energy efficient bat algorithm and a mobile sink are presented. It can choose the optimal route based on: (1) distance of a sensor node to sink, (2), bat loudness, and (3) energy level of its battery. Using the OPNET simulator version 11.5, the proposed method was simulated by bat and NODIC protocols, and IEEE 802.15.4 and the results were considered in terms of energy consumption, end to end delay, signal to noise ratio, the success probability of sending data to sink and rate of passing data. The results of the simulation showed the use of the above-mentioned parameters in the proposed method leads to the improvement of the network throughput against using the IEEE 802.15.4 protocol, Bat algorithm, and NODIC protocol.https://scj.kashanu.ac.ir/article_111447_5bfbf42b0572634eeba4192a1e9513b9.pdf