Desktop version

Home arrow Computer Science

  • Increase font
  • Decrease font


<<   CONTENTS   >>

RELATED WORKS

For solving the vehicle routing problem (VRP) in an effective fashion, a large number of heuristic approaches have been presented for logistics routes. Tabu search models as well as simulated annealing (SA) are employed by Osman [10]. The variable neighborhood search (VNS) and iterated local search (ILS) have been presented in [11]; the effective models are used to attain better solutions to enhance the primary route plan. Sze et al. [12] projected an adaptive VNS (AVNS) technique, which integrates large neighborhood search (LNS) in the form of diversification principle to use in a capable VRP. Miranda-Bront et al. [13] assumed a cluster-first route-second along with a greedy randomized adaptive search procedures (GRASP) meta-heuristic to resolve VRP. Simeonov et al. [14] developed a learning-dependent population VNS method to report the practical logistics issue motivated by a gas delivery firm in the United Kingdom. Even though VNS is a robust optimization model to withstand the complete local search, it has not been established with structured storage model.

Long-term memory infrastructure of EAs could enhance the limitations of VNS in a structured storage method. Traditional works are used in developing solutions according to EA as well as expert brilliance. A version of GA method is developed in [15] to optimize the routing distance according to practical domain of VRP. Sripriya et al. [16] presented a hybrid genetic search along with a diversity control under the application of GA to resolve VRP. The route that has minimum distance is interchanged by using a crossover and two mutation tools in this study. Zhang and Lee [17] proposed an enhancement of traditional ABC model to deal with VRP. As the EA-relied methods often require maximum duration to explore solutions, more number of techniques were used effectively like heuristic model as well as greedy techniques for solving VRP. An extended ACO integrates the semi- greedy heuristic NEH, which has been developed by Chen et al. [18] for VRP. In order to enhance, the ACO function boosts the fundamental ACO that eliminates local optimum and concatenates the adjacent searching model. Gupta and Saini [19] used the extended ACO, which comprises pheromone upgrading model and 2-opt model to increase the generalized route for VRP. Though VRP is assumed with practical limitations for logistics, the issues are regarding the mechanism of sharing few users. As people become busier, the logistics services to pick up any things from customers are considered to be more significant. Hence, managing two diverse kinds of logistics services are the severe problem. Several methods available in the literature are presented in [20-22].

The similarities of these techniques are used to combine the intelligence and knowledge from professionals at the time of iteration process to find optimal logistics solutions in an effective fashion. Mostly, such techniques are applied to explore optimized logistics solution when compared with heuristic approaches at same implementation time. But the storage as well as processing cost is a major challenge since it has massive data while the iterations are saved. The logistics industry does not provide a sufficient processing resource to operate an evolutionary model. Also, the purpose of existing works is to reduce the overall driving distance. It has limited the count of vehicles applied that influences total processing cost of logistics. These models are termed as first-delivery-last-pickup (FDLP). It is evident that the vehicle’s space could be applied effectively when a courier is filled with products from users on a back way. A set of three issues have been resolved by the HFMPSO model as given below. [1]

• Consuming quality and efficiency: Several real firms of logistics are desired to attain an optimized logistics solution.

PROPOSED METHOD

Figure 2.1 shows the overall system framework of the HFMPSO model. The proposed logistics platform is composed of few vehicles that have restricted capacity and a logistic graph contains a depot and few users are from urban area. At the primary stage, every delivery package has been gathered from a depot and each pickup package is placed at the nearby users. In HFMPSO, it is fixed with an upper bound of loading rate to select the number of packages that a vehicle can hold. HFMPSO is constrained with solution generation. It is mainly employed to explore the suitable logistics solution, which is comprised of three phases: package dividing, route scheduling, and package insertion. Hence, every package is classified as massive logistic path and a logistics solution is produced to transmit packets. As the logistics conditions as well as the needs vary in a rapid manner, it has been decided to attain the suitable logistics solution in an effective manner. On the other hand, it is evident that vehicle’s space is applied productively when a courier carries the packages from clients on a back way; hence, the upper bound is named as loading rate p, which is developed to model the way of holding a package from a depot. The loading rate p is considered as 0. It refers to a higher vehicle potential in package delivery.

Overall architecture of HFMPSO algorithm

FIGURE 2.1 Overall architecture of HFMPSO algorithm.

  • [1] A package has to be divided. In several cases, a vehicle is not capable of managingmassive package count by a wider margin. It defines that the packages should be classified into few bunches for transport facility. • Vehicle space control: When there are two types of logistics needs that are to be processed at the same time, the courier has to verify whether the vehicle has sufficientspace to carry packages at any time and place as vehicle’s ability is fine.
 
<<   CONTENTS   >>

Related topics