f. jolai

Drone delivery as a novel approach for parcel delivery has been under the focus of many scholars and practitioners. In this regard, this paper introduces a stochastic-fuzzy multi-objective optimization model for designing a last-mile delivery system with drones and ground vehicles. The first two objective functions aim to minimize the detrimental effects of the delivery system on the environment and the total costs. The last objective function maximized the system's reliability by considering the breakdown probability of both drones and ground vehicles. Then, AUGMECON2 is utilized as an exact method to solve the proposed model. besides determining the number of required drones and ground vehicles, the model indicates locations and capacities of facilities where vehicles start their one-to-one trips to meet the customer demands. The proposed model is then validated by applying it to a real case study of an e-commerce company in Karaj, Iran. The findings suggest that the system's total cost rises when the reliability increases and the environmental impacts decrease. Furthermore, when both drones and ground vehicles are considered for meeting the customer demands, the delivery system functions better in terms of costs, environmental impacts, and reliability than when only one mode of delivery is considered.

Refunding and bundling reservation are known as two popular methods to increase profit where in recent years have gained attention of researchers. One main application of refunding policy emerges for online product sale methods, where consumers can be refunded by returning goods which are not favorite according to their interest. Examining three scenarios including refunding, bundle reservation and refunding along with bundle reservation policies, we will investigate a model for each corresponding scenario. We try to compare two refund and bundle reservation pricing policies in a two-level supply chain including one manufacturer and one wholesaler, and we provide a combined model including two products. The demand is constant and also the population-related information about the division of the population into two types of consumers, strategic consumers (consumers who can predict the second stage discount) and myopic consumers (consumers who can not predict the second stage discount) are available. In addition, the percentage of consumers who refund the product due to regret, the inability to install the product or other reasons, is constant and is independent of the amount of refund. We show that the combined model is optimal and has a higher profit margin than any other policy alone.

Multi-criteria decision-making (MCDM) methods have been received considerable attention for solving problems with a set of alternatives and conflict criteria in the last decade. Previously, MCDM methods have primarily relied on the judgment and knowledge of experts for making decisions. This paper introduces a new data- and knowledge-driven MCDM method to reduce experts’ assessment dependence. The weight of the criteria is specified by using the extended data-driven DEMATEL method. Then, the ranking of alternatives is determined through knowledge-driven ELECTRE and VIKOR methods. All proposed methods for weighting and rankings are developed under grey numbers for coping with the uncertainty. Finally, the practicality and applicability of the proposed method are proved by solving an illustrative example.

In the process of hazardous material transportation, the risk is a significant factor that should be considered due to the potential severe consequence of an incident. Regardless of risks, time is a paramount concern that should be considered in hazardous material transportation. In this way, this paper introduces a bi-objective model for a vehicle routing and scheduling problem of hazardous material distribution problems under the fuzzy condition to minimize both total distribution time and risks. In the proposed model, the fuzzy inference system and fuzzy failure mode and effects analysis are applied to identify and calculate the high-level risks instead of the previous simple methods for the first time. Moreover, Jimenez method and fuzzy goal programming are respectively utilized to convert the fuzzy bi-objective model into the same crisp and single-objective one. Besides, to cope with the NP-hardness of the large-sized problems, two meta-heuristic algorithms namely invasive weeds optimization and genetic algorithm is used, and several sensitivity analyses are performed to prove the efficiency of the proposed approach. The performance of the proposed algorithms is also assessed through a comparative study. Finally, the proposed model is implemented to a real case study to prove the validity of the model.

Considering the importance of validation of customers in the cross-dock and since this is one of the problems of implementing cross-dock system in Iran, this study attempted to extract customer validation criteria. The purpose of the research is to eliminate the distrust of distributors in receiving the funds of the sent items and the statistical sample of this research is the experts of the systemofdistributionofgoodsandvalidation, indicatorswerecollectedbyusingDelphimethodand questionnaire and AHP method was used to calculate the weight and the rank of indexes.

The system of distribution of goods and services, along with other economic developments aroundthe world, is rapidly evolving. In the world of distribution of goods, the main focus is on makingdistribution operations more effective. Due to the fact that the cross-dock has the advantage ofremoving intermediaries and reducing the space required for the warehouse, it is worth considering.Among the methods of cross-docking, the post-distribution method is important in terms of uncer-tainty. Due to the importance of the issue of the post-distribution method in cross-dock, this paperaddresses the uncertainty of demand in cross-docking. For this purpose, a linear programming modelhas been developed for post-distribution cross-dock, and then solved an example by the use of themeta-heuristic whale algorithm. After that, uncertainty enters the model and the robust counterpartof the model present based on the robust optimization approach with using interval and polyhedralcollective inductive uncertainty set. The results shows the model could control the demand uncer-tainty in distance zero until 20 percent and the model does not let the changing of demand effortsconsiderably on the scheduling of the cross-docking.

Taking into account competitive markets, manufacturers attend more customer’s personalization. Accordingly, build-to-order systems have been given more attention in recent years. In these systems, the customer is a very important asset for us and has been paid less attention in the previous studies. This paper introduces a new build-to-order problem in the supply chain. This study focuses on both manufacturer's profit and customer's utility simultaneously where demand is dependent on customer's utility. The customer's utility is a behavior based upon utility function that depends on quality and price and customer's preferences. The new bi-objective non-linear problem is a multi-period, multi-product and three-echelon supply chain in order to increase manufacturer's profit and customer's utility simultaneously. Solving the complicated problem, two multi-objective meta-heuristics, namely non-dominated ranked genetic algorithm (NRGA) and non-dominated sorting genetic algorithm (NSGA-II), were used to solve the given problem. Finally, the outcomes obtained by these meta-heuristics are analyzed.

This paper develops an MILP model, named Satisfactory-Green Vehicle Routing Problem. It consists of routing a heterogeneous fleet of vehicles in order to serve a set of customers within predefined time windows. In this model in addition to the traditional objective of the VRP, both the pollution and customers’ satisfaction have been taken into account. Meanwhile, the introduced model prepares an effective dashboard for decision-makers that determines appropriate routes, the best mixed fleet, speed and idle time of vehicles. Additionally, some new factors evaluate the greening of each decision based on three criteria. This model applies piecewise linear functions (PLFs) to linearize a nonlinear fuzzy interval for incorporating customers’ satisfaction into other linear objectives. We have presented a mixed integer linear programming formulation for the S-GVRP. This model enriches managerial insights by providing trade-offs between customers’ satisfaction, total costs and emission levels. Finally, we have provided a numerical study for showing the applicability of the model.

Flow-shop scheduling problem (FSP) dealswith the scheduling of a set of n jobs that visit a set ofm machines in the same order. As the FSP is NP-hard, thereis no efficient algorithm to reach the optimal solution of theproblem. To minimize the holding, delay and setup costs oflarge permutation flow-shop scheduling problems withsequence-dependent setup times on each machine, thispaper develops a novel hybrid genetic algorithm (HGA)with three genetic operators. Proposed HGA applies amodified approach to generate a pool of initial solutions,and also uses an improved heuristic called the iterated swap < /div>procedure to improve the initial solutions. We consider themake-to-order production approach that some sequencesbetween jobs are assumed as tabu based on maximumallowable setup cost. In addition, the results are comparedto some recently developed heuristics and computationalexperimental results show that the proposed HGA performsvery competitively with respect to accuracy and efficiencyof solution.

The vehicle routing problem with backhauls (VRPB) as an extension of the classical vehicle routing prob-lem (VRP) attempts to define a set of routes which services both linehaul customers whom product are to be delivered and backhaul customers whom goods need to be collected. A primary objective for the problem usually is minimizing the total distribution cost. Most real-life problems have other objectives addition to this common primary objective. This paper describes a multi-objective model for VRPB with time windows (VRPBTW) and some new assumptions. We present a goal programming approach and a heuristic algorithm to solve the problem. Computational experiments are carried out and performance of developed methods is discussed.