Journal of Operational Research in its Applications
Year:2016 | Volume:13 | Issue:3 (50)|Papers Count:9

Clustering is the process of classifying data into groups or specific clusters based on their similarity, which is one of the commonest approaches in a number of scientific fields. In the literature on clustering, in order to avoid being tapped in local optimization, researchers have applied a number of super-innovative strategies to solve clustering problems which are originated from natural and social phenomena. In the current paper, harmonic searching algorithm (HAS) which is based on musical instruments have been applied as a capable algorithm in solving clustering problems. HSA is presented to evaluate the capability of a number of standard datasets. The results of the simulation indicated that this algorithm has high efficiency in finding optimal solutions.

Islamic Azad University as a private academic system with centralized decision making environment in Iran aiming to reengineer its structure in 2 recent years in such a way that efficiency of provincial units and the whole university get increased. This research aims to develop data envelopment analysis models to measure the efficiency of provincial units of this university as decision making units in multi-stage systems with parallel network structure in a case study. Despite the development of DEA models in the last decade, a model for measuring efficiency of parallel network structure considering undesirable outputs and uncontrollable inputs has not been developed yet. In order to develop such a model, the provincial units of Islamic Azad University have been studied with regard to the horizon 1404 agenda. Using data envelopment analysis without output/input orientation and with preference structure, a performance measurement model of decision making units and whole of the university has been developed in this study.

Uncertainty is one of the intrinsic characteristics of natural phenomena, and optimization is not an exception here. Parameters in an optimization problem may not be known exactly or may experience variation after finding an optimal solution. In this paper we consider the linear optimization problem in an uncertain environment. It is assumed that the problem parameters are provided by an expert in the domain. Uncertainty Theory initiated by Liu in 2007, an axiomatic framework for modeling of human reasoning, is the underlying idea of the approach here. Two different uncertain linear programs are presented in this study; moreover, feasibility and optimality of the models as well as relations between these two models are investigated. Obtained results are depicted in two illustrative simple examples.

In this current research, job shop scheduling problem by considering a parallel assembly stage and lot streaming has been studied to minimize the completion time of products. At first, a mixed integer linear programming model has been introduced to describe the problem. To confirm the accuracy of the model, some small size instances have been solved with the use of GAMS. Since this problem is classified as NP-hard, the hybrid algorithms based on genetic algorithm, simulated annealing (SA), variable neighborhood search (VNS) and parallel variable neighborhood search (PVNS) have been applied to solve the problem in medium and large dimensions. In order to verify the effectiveness of the proposed algorithms, relative percentage deviation (RPD) along with statistical analysis have been presented. Computational results showed that HGAPVNS algorithm can outperform the other proposed algorithms in terms of solution quality.

Banks managers need to evaluate different branches' performance in order to plan and organize their branches. Additionally, selecting an appropriate location for opening one or more main branches is more important in order for the opened main branch to be able to cover predefined branches in the radius or non-radius distance. The concept of facility efficiency as defined by data envelopment analysis (DEA) is presented as another location objective to investigate the effect of facility location on their efficiencies for demand servicing on problem. Many types of the location allocation problems have been developed to find optimal spatial patterns with respect to criteria such as cost, time, coverage and access. In this paper, this method has been developed in locating Mellat bank main branches. The performance of the combining model is investigated with an example. By combining data envelopment analysis with locating Mellat bank main branches, the costs due to non-covering the definite numbers of demand locations is minimized, and also the main branches efficiency score obtained based on data envelopment analysis technique to find optimal and efficient facility location is maximized. Solving the DEA efficiency measure, simultaneously with other location modeling objectives, provides a promising rich approach to multi-objective location problems. Results are obtained and analyzed by using the GAMS software.

In the current study, we study the evolutionary dynamicity of a population, composed of two kinds of individuals namely as type F (Fight) and type Y (Yield) groups which are also distributed randomly. In this model of game, individuals are the same as economic agents interacting in pairs and exchanging their money in a random way. Of course, individuals have a fixed behavior during the game. In this game also, evolution takes place in different periods of time between individuals frequently. We present a continuous nonlinear operator, which describes and fulfills this discrete time evolutionary game. Furthermore, we set up the adequate mathematical framework to obtain the fixed point of this operator. Based on this equilibrium state of the evolutionary operator, it is proved that the possessions of the individuals vanish with time.

The decision to select and allocate orders to suppliers in the supply chain is considered as one of the most discussed topics in this chain and includes many factors that are sometimes in conflict with each other. In this paper, confirmatory factor analysis and analytic network process is used to rank suppliers; in addition, integer goal programming method is used for targeting all the tangible and intangible factors as well as quantitative and qualitative amounts of supply chain levels. Reducing the purchasing and transportation cost, reducing the waiting time to receive the order, increasing product quality and so forth are considered as the goal objective; moreover, in this study, company policy based on reduction of the numbers of suppliers and policy of order amount to suppliers were evaluated and ordering to company's suppliers has been done based on company's model. Finally, a sensitive analysis has been conducted on the main limitations of the company.

Data Envelopment Analysis (DEA) is a powerful technique for measuring the relative efficiency of a group of decision making units (DMUs) with multiple inputs and outputs. In real world, most input and output data are imprecise or fuzzy; while conventional DEA models use crisp DMUs. In this paper, we aim to propose a model to evaluate the performance of DMUs with fuzzy values. On the other hand, since in radial DEA models, only changes relative to inputs or outputs are considered, slack values are ignored in evaluating the performance of units and this is a flaw in these types of models. Moreover, non-radial models are known as the best estimators of efficiency in DEA models. Therefore, we were persuaded to use a non-radial model with fuzzy data to evaluate the performance of DMUs. Finally, a numeric example is presented to show the importance and characteristics of the proposed model.