چكيده لاتين
Abstract
This research investigate a combined problem under the title of 3L-HFFCVRP problem by considering a fixed (limited) fleet of heterogeneous vehicles, each of them has its own dimensions, fixed (usage) costs, variable (travel) costs and stopping costs, and also by defining the practical constraints for three-dimensional loading, including the limitation in the possible directions of the boxes, the minimum required base support of boxes for the vertical (static) stability and the fragility of boxes (or ability of bearing the load on themselves). In fact, the problem investigated in this research is a combination of a type of vehicle routing problem (VRP) with a type of container loading problem (CLP), which is often seen in those manufacturing companies that produce various products and pack them in cubic boxes with various weights and dimensions and send them through a heterogeneous fleet of vehicles to their customer destination in order to satisfy their demands. It is obvious that due to the high cost of transportation, for such companies, in addition to the necessity of optimizing the routes of vehicles, it is very important to design the suitable layouts for loading goods in containers,
At first, a pure integer linear programming model is proposed to solve the vehicle routing problem with three simultaneous characteristics of fixed heterogeneous fleet, multi-products and split delivery called HFFVRP/MPSD. This model minimizes the total cost of transportation (i.e., the sum of fixed costs, travel costs, and stopping costs) by selecting a number of vehicles from the fleet of vehicles and determines the products that should be loaded in each of them, as well as specifying the order of the customersʹ visits by each vehicle and determining which items each vehicle delivers to which customer. In this model, special constraints have been used to break the sub-tours, which, while reducing the number of inequality constraints compared to the conventional method, can also determine the order of meeting customers. The proposed exact method obtains feasible solutions for all tested sample of problems (with up to 50 customers), but since it is not able to achieve the optimal solution in an acceptable time (up to 3600 seconds) for problems with more than 10 customers, and it takes a lot of time to achieve satisfactory solutions for problems with more than 15 customers; in order to improve the speed of solving the problem and increase the efficiency of the model, a meta-heuristic model based on the hybrid genetic algorithm, has been developed. The presented meta-heuristic model has been able to obtain higher quality solutions for the problem by reducing the execution time by more than 95%. Also, the improvement of the quality of the solutions obtained for problems with more than 10 customers was more than 45% on average. Finally, in order to solve the combinatorial problem, by introducing three-dimensional loading constraints to the aforementioned meta-heuristic model, a complete model based on the hybrid genetic algorithm has been developed for the 3L-HFFCVRP problem.
