Journal of Industrial and Systems Engineering
Volume:15 Issue: 1, Winter 2023

Economical, environmental, and social issues are significant challenges for industries and governments nowadays. The spare parts impose high inventory costs on the companies and require human resources, energy, and budget for the repair operations. These issues justify integrating repair, and inventory management decisions to reduce costs. Since the system is interacting with the environment, incorporating the sustainability dimensions with network design and planning decisions help managers to make more reliable decisions. We investigated the social and environmental dimensions to cover the sustainability dimensions of the spare part supply chain. These attributes contribute to industry-oriented properties in real-world problems. This paper investigates a multi-objective model to minimize costs while maximizing sustainability in a repairable spare part supply chain. Life cycle assessment (LCA) is utilized to assess social and environmental dimensions. Finally, the model is solved using NSGA-II with a priority-based encoding and decoding procedure. The findings shed light on contributing to formulating the spare part supply chain sustainability which integrates the network design and planning decisions resulting in more reliable outcomes.

In present study, a mathematical model for designing a humanitarian supply chain network and vehicle routing problem considering cross-dock is proposed where a Non-dominated Sorting Genetic Algorithm (NSGAIII) is used for implementing the proposed model in a large-scale problem. Since the model was implemented in a large-scale case, various sensitivity analyses were performed to extract the results. Hence, the results showed that the costs have more effect on the first objective function (patients compared to total injuries) and the second one (shortage), respectively. Compared to the other two objective functions, the impact on the cost function is negligible. The effect of transportation cost of relief goods/supplies from the supplier to the warehouse on the first objective function is higher than the others; however, the effect of this cost is further than that of the cost from the supplier to the distributor, accordingly, in comparison to the previous cost, the output reacted more to this cost. The transportation cost from the distributor to the warehouse (cross-docking) has less effect on the cost function unlike the transportation cost from the supplier to the warehouse. Nevertheless, the result shows that an increase in the cost can lead to a considerable increase in the ratio of patients to total injuries as well as shortage. In other words, the objective functions would deteriorate when this parameter tends to be increased.

Parametric models are considered the widespread methods for time series forecasting. Non-parametric or machine learning methods have significantly replaced statistical methods in recent years. In this study, we develop a novel two-level clustering algorithm to forecast short-length time series datasets using a multi-step approach, including clustering, sliding window, and MLP neural network. In first-level clustering, the time series dataset in the training part is clustered. Then, we made a long time series by concatenating the existing time series in each cluster in the first level. After that, using a sliding window, every long-time series created in the previous step is restructured to equal-size sub-series and clustered in the second level. Applying an MLP network, a model has been fitted to final clusters. Finally, the test data distance is calculated with the center of the final cluster, selecting the nearest distance, and using the fitted model in that cluster, the final forecasting is done. Using the WAPE index, we compare the one-level clustering algorithm in the literature regarding the mean of answers and the best answer in a ten-time run. The results reveal that the algorithm could increase the WAPE index value in terms of the mean and the best solution by 8.78% and 5.24%, respectively. Also, comparing the standard deviation of different runs shows that the proposed algorithm could be further stabilized with a 3.24 decline in this index. This novel study proposed a two-level clustering for forecasting short-length time series datasets, improving the accuracy and stability of time series forecasting.

Considering the extensive application of dynamic multi-objective optimization problems (DMOPs) and the significance of the quality of solutions, developing optimization methods to find the finest solutions takes a privileged position, attracting considerable interest. Most optimization methods involve multiple conflicting objectives that change over time. The present article develops an electromagnetic field optimization (EFO) using decomposition, crowding distance, and the quantum behavior of particles techniques to solve multi-objective problems. In the proposed algorithm, the position of new particles is determined between the neighbors within the MOEA/D by drawing inspiration from the quantum delta potential well model, the nonlinear trajectory of quantum-behaved particles, and the interactions of electromagnetic particles introduced from positive and negative fields, which can offer superior exploration and exploitation. To develop the proposed algorithm for solving dynamic problems, the mean difference between particles' center of mass in the two latest changes to predict the extent of change is applied along with polynomial mutation and random reproduction. A total of 9 benchmarks from the set of DF functions and two metrics, i.e., MIGD and MHV, are used to assess the performance of the proposed algorithm. The results from 20 independent runs of the proposed algorithm on each benchmark function are compared with the results from other algorithms. The Wilcoxon Rank-Sum non-parametric statistical test is applied at the significance level of 5% to compare the mean results. The experimental results indicated that the proposed algorithm gains a significant superiority in metrics MIGA and MHV in most experiments. The simultaneously great results of these two metrics indicate a superior distribution and approximation of the Pareto front.

Optimal time interval between inspections of the redundant systems is raised as a substantial issue to plan a preventive maintenance model for maintenance planners. To have an optimal time period for preventive maintenance of systems, especially complex systems such as redundant systems, two variables of maintenance are mostly connived. Against other studies in the literature of preventive maintenance in which repair time is a negligible factor as an assumption, repair time is considered as a noticeable variable incorporated into the model developed in this paper. Another contribution, the number of facilities, is focused as a significant variable used in real applications. Particularly, systems with complex performance needing technical repair facilities (i.e., technical repairmen, tools, materials, outsourced repair, etc.). In this regard, parallel systems have been analyzed stochastically using the way of preventive maintenance in which repair time would be contemplated as an essential factor in maintenance planning. Using Markov chain, a model based on expected total cost per time is made to demonstrate that a proper time interval achieving lowest possible cost is obtained by taking into account repair time and the number of repair facility. Three models are studied as instances of redundant systems to find the optimal time interval between inspections. These models differ in the number of repair facilities (i.e., one, two and three repair facilities). A sensitivity analysis is done to depict the variability of input variables over optimal the expected total cost per time and time interval between inspections. As a main contribution, the repair time could be an essential factor in maintenance planning, this study contemplates this factor in redundant systems.

In today's world, due to the competitive nature of the market and the lack of certainty in the amount of order and also the time of ordering products, it has led to the effective response of sales centers to customers is not done properly. This is due to the lack of proper location of distribution and sales centers and optimal allocation of customers to each center. Therefore, considering the importance of locating distribution centers, in this article, the issue of locating distribution centers of e-shops in conditions of uncertainty has been developed. The main purpose is to provide a model for profit maximization and minimization of the total transfer time of electronic products between distribution centers and customer clusters. To examine the developed model, three different problem solving methods have been considered, including the Epsilon constraint method, the NSGA II algorithm and the MOPSO. The results obtained from the analysis of the sample problem in small size show that NSGA II algorithm has 14 efficient answers, MOPSO algorithm has 10 efficient answers and Epsilon method has obtained a limit of 8 efficient answers. The computational results show the high efficiency of the MOPSO algorithm in obtaining the optimal weight of 0.9744 in solving large size problems.

The perishable dairy industry has to deal with multiple challenges such as demand forecasting, price fluctuations, lead time, and inflated orders along with difficulties of climatic and traffic conditions, storage areas and shipment in unfavorable circumstances. This research introduces a robust bi-level mathematical model to optimize a multi-echelon Perishable Supply Chain (PSC. To this end, integrated multi-objective Mixed Integer Linear Programming (MILP) models are developed to formulate the problem. stochastic deterioration rate is taken into account as the main factor that determines model performance due to perishability of products. In order to contribute to the literature, mainly by addressing uncertainty and perishability, a solution technique based on robust programming and -constrait approach is developed to accommodate suggested bi-level model. This technique can deal with problem uncertainty while also ensuring the robustness of the overall system. Sensitivity analysis is implemented along with three well-known quality indicators to assess the performance of the proposed solution method and quality of obtained solutions. Finally, real case study is provided using the CPLEX solver to showcase the applicability of the proposed methodology and discuss the complexity of the model. Results demonstrate the efficiency of the proposed methodology in finding optimal solutions.

In today’s turbulent world, unpredictable events with various effects on human health have highlighted the importance of an agile and integrated network to provide health services. Providing health services to applicants is realized when procurement of the required health items is effectively and efficiently managed. In the present research, a practical approach is taken to design an integrated network to provide laboratory services and manage the procurement of related items under uncertainty. For this purpose, a multi-product and multi-period optimization model is presented to minimize the expected costs. Also, a scenario-based robust optimization approach is adopted for uncertainty programming. Moreover, a real-world case study in Iran is employed to ensure the effectiveness and efficiency of the presented model. Integrating the network of providing laboratory services and managing the procurement of related consumable items, employing geological software such as Arc GIS to locate potential facilities in laboratory service network, and simultaneously dealing with disruptions and operational risks in healthcare networks are the distinctive research contributions. The obtained results indicate the advantages of designing an integrated network to provide laboratory services and manage the procurement of relevant items to save costs and improve the quality of providing service to applicants. In general, it could be observed that a lack of planning to deal with disruptive incidents could cause severe damage to the performance of the studied integrated network.

Identifying and evaluating supply chain risks is one of the most challenging issues related to supply chain risk management (SCRM). Many risks may threaten a supply chain, but upon the costs, managers had better pay attention to those with the highest impact. The paper advances to identify and rank Information and Communication Technology (ICT) supply chain risks and investigate their intereffects in a directed graph through the social network analysis approach and experts' opinions. Firstly, ICT supply chain risks were determined based on semi-structured interviews with organizational experts on the viable system model (V.S.M.). Then, they were asked to set a score between zero and five based on the impact of each risk on the other risks to assigning appropriate weight to edges. Finally, ICT supply chain risks were ranked based on centrality measures. The findings indicate that social and political conditions affect the ICT supply chain. As well as, the accuracy of the information and the emergence of new technologies are other factors that have the most significant impact on additional risks in the supply chain. We also situated the analysis on Tehran Internet Holding, a large company representative sales and after-sales service agent of Iran's most outstanding digital operator.

The origin of any project is identify in its requirement and this means value creation of a project, which should be measured according to how these needs, are satisfied. At the same time, it is important, as how efficient is a project and at what price can these needs be met? In its simplest form, value is defined as the relationship between expanses and benefit, or a measure of what one loses compared to what one gains. In this article, the data needed to answer the research questions collected using the questionnaire’s tool, to analyze the collected data Descriptive statistics and inferential statistics methods utilized. In the descriptive analysis of this research from frequency table, and central and dispersion indices used in the inferential analysis part of the research, Friedman's ranking test also used.To answer the research questions firstly, we reviewed the articles and reports, also consultation with five experts from the rail transportation department thus, the main issues related to the creation of value and risk in the rail transport sector have been identified then, based on these issues, the questionnaire form compiled and distributed among 175 related employees of MAPNA’s rail transport companies. The employee’s replies to the questions asked through the questionnaire form have been coded and analyzed in SPSS software format.The findings of the research show that there are different solutions to manage and control the identified risks; these specified solutions and strategies presented in the final part of this study.

This study aims to design a humanitarian logistics network for location-routing equipped with drone-enabled delivery systems under uncertainty conditions. Here, we divided the model into two phases including pre- and post-disaster. There is an important question in pre-disaster phase: Where the central warehouses better perform to minimize the cost and time? To this end, the logistics problem represented by the transportation of relief products was modeled as a Multi Echelon Multiple Depot Vehicle Routing Problem (MEMDVRP). For solving this mathematical problem, the presented model was initially solved using meta-heuristic algorithm of Non-dominated Sorting Genetic Algorithm III (NSGAIII) in large dimensions and sensitivity analysis was performed on its effective parameters via MATLAB software. Due to the scenario nature of the problem, 4 scenarios were considered in the model and then were compared separately for each goal. Given the results, scenario 4 showed the best situation in terms of benefits maximization. Regarding the cost, scenario 4 shows the worst status and the scenarios 1 and 2 revealed the best status. It should be noted that due to the nature of cost minimization objective, the lower this value, the better the result, indicating the best cost situation for Scenarios 1 and 2 and the worst for Scenario 4. In terms of time, Scenario 4 indicated the worst condition likewise the cost. Interestingly, regarding the benefits, the Scenario 4 leads to the most benefits, so it can be said that in this scenario, as the benefits increase, the cost and time also increase, suggesting a conflict in objectives.

The increase in population, demand growth, limitation of water resources, and huge costs of water supply with the execution of new plans for water resources development have attracted more attention to the management of the existing resources and facilities exploitation. For this purpose, the present study uses optimization methods in urban water supply system programming and introduces state and flow variables to provide the urban water supply model and evaluate the factors affecting the urban water supply cycle in Guilan province using the system dynamics. Finally, a model is presented in line with the management status of the Guilan Water and Wastewater Company in Rasht branch. Evaluating the accuracy and validity of the model presented in the water supply system from different water resources shows that increasing the treatment capacity and water resources in the province can affect the treatment cost and reduce the shortage and wastewater. Regarding the parameters which have a positive effect on the amount of input water, it is required to consider appropriate systems to control the input water and manage such valuable resources. Eventually, forecasting the amount of shortage in the studied area during the next 100 years indicates a linear trend that the number of shortages increases in an upward manner in each period due to the increased population and decreased amount of precipitation.

In recent years, the outbreak of COVID-19 has led to burnout of healthcare personnel. Accordingly, more attention should be paid to nurses scheduling and their preferences. The Nurse Scheduling Problem (NSP) as an optimization concept provides suitable nurses' schedules by focusing on the system requirements. In this study, a new NSP is developed in which the factors and consequences of nurses' burnout are considered simultaneously. In the proposed model, new constraints are formulated to define the undesirable shifts. Due to the seniority rules, it is tried to restrict the numeral of these shifts in the generated timetable to improve the burnout of nurses. In addition, an attempt is made to fairly allocate the requested leave of nurses by considering their leave days during the previous horizons. In the presented model, the timetable of nurses is flexible to cope with the absence of employees, and the required personnel are covered by changing shifts among nurses. To solve the developed problem, Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) are coded, and their results are compared with GAMS in different test problems. Taguchi method has been applied for parameter tuning of these algorithms. The final results prove that the GA outperforms PSO both in obtained solution and CPU time. GA has only a 0.22 % optimality gap on average. Finally, the proposed model is implemented in an actual case study in Iran. The generated timetable improves nurses' performance and the level of medical services by controlling the causes and consequences of burnout.

The key to solving the problem of obtaining complex facilities is to create a suitable credit rating model that can provide technical support for the approval of granting facilities provided by small and micro enterprises. Credit rating agencies perform assessment to support financial institutions in processing debts. Added literature in the field of credit rating from January 2015 to August 2023 was analyzed to discover opportunities for further research. Bibliometric analysis was used to understand the existing literature. Subsequently, through structured review theories, the methods used by researchers and credit rating agencies were examined. A hybrid literature review was developed by integrating bibliometric and structured review of research articles from widely recognized databases. A sample of 72 articles has been made and studied to identify the gaps in the field of credit rating and create a suitable solution to fill such gaps. The results showed that most studies appeared as post-financial crisis effects reported in 2016 and 2023. It contributes to the existing literature by encouraging researchers and credit rating agencies to develop a specific credit rating system by evaluating existing models and improvising them by adopting advanced techniques such as multiple regression, neural networks, aggregate learning, and machine learning.

This paper studies a name-your-own-price (NYOP) mechanism in which the retailer allows buyers to participate in the pricing process by submitting bids. Buyers can place both joint and individual bids to purchase products either as a bundle or individually. The retailer utilizes NYOP and posted-price channels simultaneously. The focus of this paper is to assess the impact of adding the postedprice channel and bundling option on buyer behavior and retailer profit. The paper develops a two-stage model where the first stage involves the buyer’s decision on participating in NYOP. Moreover, buyers can choose between bidding for a bundle or a single item. Decisions in the second stage depend on the outcome of the first stage. Four distinct purchasing scenarios are formulated to outline the potential ways that buyers can use to purchase products. Furthermore, the buyers’ learning effect on their bidding strategy is considered. A dynamic programming approach with backward induction is employed to solve the problem. Moreover, the concavity analysis is used to obtain the solution of each nonlinear subproblem. Then, a solution algorithm based on mathematical analysis is proposed. Results reveal that the frictional costs of the first period have a greater impact on the buyer utility than those of the second period. Moreover, applying the NYOP alongside the posted-price can enhance the retailer’s profit. In particular, the retailer can use the NYOP and bundling mechanisms as encouraging tools to attract buyers and increase his profit. Thus, NYOP is a very effective instrument for market penetration.