Research on Location-Routing Problem with Empirical Analysis for Regional Logistics Distribution

The location of the distribution facilities and the routing of the vehicles from these facilities are interdependent in many distribution systems. Such a concept recognizes the interdependence;attempts to integrate these two decisions have been limited. Multi-objective location-routing problem (MLRP) is combined with the facility location and the vehicle routing decision and satisfied the different objectives. Due to the problem complexity, simultaneous solution methods are limited, which are given in different objectives with conflicts in functions satisfied. Two kinds of optimal mathematical models are proposed for the solution of MLRP. Three methods have been emphatically developed for MLRP. MGA architecture makes it possible to search the solution space efficiently, which provides a path for searching the solution with two-objective LRP. At last the practical proof is given by random analysis for regional distribution with nine cities.

Designing a locating-routing three-echelon supply chain network under uncertainty

Purpose-In the present research,location and routing problems,as well as the supply chain,which includes manufacturers,distributor candidate sites and retailers,are explored.The goal of addressing the issue is to reduce delivery times and systemcosts for retailers so that routing and distributor locationmay be determined.Design/methodology/approach-By adding certain unique criteria and limits,the issue becomes more realistic.Customers expect simultaneous deliveries and pickups,and retail service start times have soft and hard time windows.Transportation expenses,noncompliance with the soft time window,distributor construction,vehicle purchase or leasing,andmanufacturing costs are all part of the system costs.The problem’s conceptual model is developed and modeled first,and then General Algebraic Modeling System software(GAMS)and Multiple Objective Particle Swarm Optimization(MOPSO)and non-dominated sorting genetic algorithm II(NSGAII)algorithms are used to solve it in small dimensions.Findings-According to the mathematical model’s solution,the average error of the two suggested methods,in contrast to the exact answer,is less than 0.7%.In addition,the performance of algorithms in terms of deviation from the GAMS exact solution is pretty satisfactory,with a divergence of 0.4%for the biggest problem(N 5100).As a result,NSGAII is shown to be superior to MOSPSO.Research limitations/implications-Since this paper deals with two bi-objective models,the priorities of decision-makers in selecting the best solution were not taken into account,and each of the objective functions was given an equal weight based on the weighting procedures.The model has not been compared or studied in both robust and deterministic modes.This is because,with the exception of the variable that indicates traffic mode uncertainty,all variables are deterministic,and the uncertainty character of demand in each level of the supply chain is ignored.Practical implications-The suggested model’s conclusions are useful for any group of decision-makers concerned with optimizing production patterns at any level.The employment of a diverse fleet of delivery vehicles,as well as the use of stochastic optimization techniques to define the time windows,demonstrates how successful distribution networks are in lowering operational costs.Originality/value-According to a multi-objective model in a three-echelon supply chain,this research fills in the gaps in the link between routing and location choices in a realistic manner,taking into account the actual restrictions of a distribution network.The model may reduce the uncertainty in vehicle performance while choosing a refueling strategy or dealing with diverse traffic scenarios,bringing it closer to certainty.In addition,two modified MOPSO and NSGA-II algorithms are presented for solving the model,with the results compared to the exact GAMS approach for medium-and small-sized problems.