نشریه محاسبات نرم
سال هفتم شماره 2 (پیاپی 14، پاییز و زمستان 1397)

The SIP protocol was standardized by the IETF at the application layer for initiating, managing, and terminating multimedia sessions and has been widely used as the main signaling protocol on both the Internet and VoIP networks. Most challenges in this protocol are overload and lack of proper state distribution. These challenges cause a wide range of next-generation network users to face a sharp decline in service quality. In this article, we define the state distribution problem between several nodes where the state maintenance is accompanied by considerable consumption of resources leading to overload. For the problem solution, the goal is to increase the overall throughput of calls and the availability of servers. First, we provide a framework based on software-defined networking technology, and then we formulate the problem as an optimization problem and implement and evaluate it as a module on the proposed controller. This leads to a more scalable SIP network that dynamically determines the number of SIP requests for which the server is modeled while delegating the state maintenance to its downstream server for the rest of the requests. This is in contrast to existing SIP servers because they are statically configured to be either stateless or stateful, resulting in non-optimal call throughput. Performance evaluation is performed at two levels of infrastructure and control and the results are presented.

In the electronics industry, scaling and optimization is final goal. But, according to ITRS predictions, silicon as basic material for semiconductors, is facing physical limitation and approaching the end of the path. Therefore, researchers are looking for the silicon replacement. Until now, carbon and its allotrope, graphene, look to be viable candidates. Among different circuits, IO block is a needed ingredient for electronic systems and needs to be re-designed and optimized. In this paper, goal is feasibility analysis and design of IO blocks using graphene field effect transistors. Using these transistors, each ingredient is designed, simulated and analyzed using the HSPICE tool. Then, these ingredients are combined together and the graphene-based IO block is implemented. Similar to graphene IO, silicon- based IO block is also designed and results are compared. It indicated that propagation delay is 299.9ps with 10% edge roughness, which is 32% faster compared to silicon counterpart.

Any scientific activity requires awareness of previous related activities. Citation networks are the networks in which each document is compared as a link of a chain with its previous and next documents, and the documents with the highest number of citations are considered as the most effective ones in a domain. Most of the introduced methods use direct citations for valuing the documents. One of the challenges in ranking documents are indirect citations and determining effective features to compute the ranks. The presented method uses not only direct but also indirect citations for the valuing process. In this research, several measures for analyzing citation networks are introduced. The combination of these measures and using a data fusion method will improve the way by which documents are valued and ranked. In the presented method, an ordered weighted averaging fusion method is used to determine values based on the direct and indirect citations computed by some defined measures. Through some experiments, the presented method has been compared with some other fusion methods, which results indicate the effectiveness of our method.

Identifying the appropriate software components in the software design phase is a vital task in the field of software engineering and is considered as an important way to increase the software maintenance capability. Nowadays, many methods for identifying components such as graph partitioning and clustering are presented, but most of these methods are based on expert opinion and have poor accuracy. One of the reasons for inaccuracy of the component identification methods is the lack of attention to the fact that there are contradictions between the criteria used to recognizing a component, which to be compromised during the identification process. In this paper, a novel method based on non-dominated sorting genetic algorithm (NSGAII) is proposed to map the software component detection problem into a multi-objective optimization one. The proposed method uses the criteria of cohesions, coupling and complexity in order to identify the appropriate components. In this paper, a real study system (customers club) has been used to evaluate the proposed method. The results of the evaluation revealed that the use of the proposed multi-objective algorithm has been able to perform better than previous single-objective methods.

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 mobile sink can increase the efficiency of wireless sensor networks. It moves in a monitored environment and collects the network nodes information. Thus, by the sink we can balance the power consumption and increases the network lifetime. Determining path of the sink's movement is usually modeled as an optimization problem where finding optimal solutions require collecting value of all the network parameters at each moment and then solving the optimization model. Therefore, the optimization model requires a large amount of information processing and transmission, and consequently it is impractical and almost impossible. This paper proposes a decentralized algorithm for determining the itinerary planning of the sink based on the local information collected from the vicinity of the sink. The proposed algorithm decides based on the estimation of the remaining network lifetime and energy consumption around each candidate. The use of fuzzy TOPSIS in presenting a decentralized algorithm for planning the itinerary in wireless sensor networks is the idea proposed and implemented just in this paper. The simulation results show that the proposed algorithm can significantly improve the network lifetime.