نشریه پژوهشهای مهندسی صنایع در سیستمهای تولید
پیاپی 5 (بهار و تابستان 1394)

Nonparametric control charts are used for the controlling of processes with non-normal or unknown distributions. Also fuzzy control charts are used when there is uncertainty in the sampling data. The signed rank nonparametric control chart use for the detecting changes in mean (median). In this paper the signed rank control chart is developed with fuzzy data and the proposed chart is called as fuzzy signed rank control chart. For this purpose، three different approaches with different ambiguity are proposed. A simulation program is used for calculating of the performance of fuzzy signed rank control chart based on average run length (ARL). The control chart performance study for three proposed approaches with three different distributions as short-tailed (Uniform)، middle-tailed (Normal) and heavy-tailed (Laplace). Numerical results show suitable performance of the fuzzy signed rank control chart in the controlling of fuzzy non-normal process.

In classic inventory control system، purchasing cost must be paid when purchased items are received. Furthermore، in the mentioned models، items may have infinite lifetime، while in the real world، we have some items deteriorate over the time. Also، in order to increase the sales، the suppliers may allow the retailers to pay the purchasing cost some times after receiving the ordered products. Moreover، in the real life cases، the supplier may visit the retailer and send the ordered quantity at random time and the retailer faces to stochastic lead time; then، in this case، the retailer may face to shortage. In this paper، we will extend the periodic inventory control model under delay in payment، stochastic visit interval and partial backordering for a deteriorating item. Under general probability distribution function between replenishment epochs، we show the concavity of the expected profit function and give the condition that must hold for the optimal replenish-up-to-level in order to maximize the profit. In order to show the applicability of the proposed model، the numerical examples and sensitivity analysis are provided.

Supply chain network design is an essential subject to achieve a competitive advantage in today’s market. In addition، recycling and recovering of products because of environmental aspects and decreasing of row material consumption is a vital subject. This paper proposes a design of closed-loop supply chain network with considering of multi-part collection centers under uncertainty to obtain minimum environmental damage as well as decreasing total costs. Because of uncertainty nature of related data in the real world، it is essential to use uncertainty approaches in the proposed network. In this regard، we suppose that demand and recovery are defined in stochastic environment. In reality، we apply a heuristic algorithm named as Sample average approximation algorithm with scenario based. By increasing the size of problems، the proposed hybrid genetic-SAA algorithm is used. Validity of the proposed model is illustrated through numerical examples. Results show that with consideration of multi-part collection centers، not only recovered products and recycling saving costs are improved but also، customers can receive higher quality products with satisfying of government because of environmental damage reduction.

This paper is dealing with Two-dimensional loading time-dependent vehicle routing problem. A new mathematical model is proposed and solved. Aforementioned problem is about delivering rectangular items to customers. In the problem that we considered، travel time between two nodes depends not only on their distance، but also depends on departure time from origin node. Such an assumption seems to be important for route design in urban areas، because traffic jam changes travel time on beginning and ending of work time. Despite applicability of such an issue، there is not any research considering this problem. In this paper، we proposed a new mathematical model. For evaluating and validating this model، some small-scale problems solved and for large-scale problems، a simulated annealing and an improved genetic algorithm are proposed. For checking feasibility of loading of assigned items to a vehicle، a collection of heuristic algorithms is used. Computational results confirm the effectiveness of the solving approaches.

Hub location problems have important role in transportation networks and telecommunication systems. These problems have received much attention in the recent decades. In p-hub median problem، hub facilities are located in a network and non-hub nodes are allocated to hub nodes such that the total routing cost is minimized. In this paper، the r-Allocation p-Hub median problem has been studied. In this problem، each node can be allocated to at most r hubs. The purpose of this paper is to study and design of an exact solution approach based on benders decomposition to tackle the proposed problem. Based on problem structure، we proposed some strategies to accelerate the convergence of benders decomposition algorithm. The computational results of the proposed exact algorithm are presented using the Turkish network and USA423 data sets. Computational experiments confirm the effectiveness of the proposed method

Nowadays، reliability problem is one of the most important issues at early stages of planning، designing and control of production systems. In this paper، the authors try to develop a hybrid simulation-optimization method in order to solve reliability problem for series-parallel system in uncertainty condition with stochastic constraints. We propose a method in which first، a population of fairly good solutions is generated by simulation technique and then، these solutions are improved by Particle Swarm Optimization (PSO). To assess the efficiency of the proposed algorithm، the results of this research were compared with previous similar studies. The results show that the proposed algorithm is superior. The strength of the proposed algorithm is higher average and lower standard deviation of solutions obtained in successive iteration.

The location routing problem (LRP) is presented with the aim of specifying both routing and location decision simultaneously and coordinate these two matters is very important in designing distribution networks of supply chain. This importance is for that the suitable coordinating between location and routing has a powerfuul affect on supply chain performance in distribution systems and also can improve it’s efficiency indexes. Although for simplifying، these two matters are usually analyzed and solved in two separated phase، but this would cause to lost the ideal benefit and global optimum solution. In this paper، the routing and location problem with considering real word conditions and restrictions like diversity of vehicles and restrictions of some vehicles movement in specified route are investigated concurrently. After presenting summery of previous research، we investigate the problem that mentioned earlier and relevant variables and parameters will be declared. Then mathematical model will be extended. In addition، mathematical model with modeling software GAMS will be implemented. With assuming this problem is kind of NP-hard problem، thus we solve this problem in small scale. In order to solve the considered problem in larg scale، a model will be presented based on ant colony. So with our specific parameters، diverse testing problems will be designed and outcome of these problems will be analyzed to show the algorithm efficiency.

Network maximum flow problem trying to find the maximum flow that can transmit from the source nodes to sink nodes. The purpose of this research is to improve and simplify the initial graph that given as a problem to solve by network maximum flow algorithm. Most of maximum flow algorithms solve the problem with finding route & transmitting based on basic feasible solution. In this paper، we attempt to present a new heuristic approach according to depth concept in graph to reduce the problem time complexity with removing arcs and nodes. This algorithm is compatible with multi-source multi-sink algorithms in the literature. Finally، to demonstrate applicability of the proposed approach، several test problems are first generated; then، the comparisons are carried out. The results show that the proposed heuristic can simplify the problem and consequently reduce the complexity of maximum flow problem