A Distribution Planning Problem Based On Separated Factory And Warehouse

As the increasing shortage of land resources, factories in big cities tend to move thewarehouse to countryside. In the meantime, products should be delivered to customersinstead of warehouse once it has been produced in order to reduce inventory and be quickrespond to customers. For the limited space in the product lines, factories need to deliverthe product as quickly as possible to avoid products being beyond the storage space onthe one side, it expects the full load delivery to reduce transportation cost on the otherside. In this situation, it is important to optimize transportation plan under the conditionof separation of factory and warehouse. It is also the key factor for the factory to decreasethe cost of logistics.This article studies optimizing products transportation in three-echelon production-distribution network including one factory, one warehouse, multiple customers. In thisnetwork, the rule is described as followed: the factory and the warehouse are separated,so there is no warehouse in factory to store the products; all the shipment in this networkis one to one; the daily production plan of the factory is known; weekly customersrequirement are known but daily requirement is unknown. The aim of the optimization isto minimize the cost of transportation in production-distribution network and formulatethe arrival time of every vehicle. Therefore, the key of optimization is as follow: wherewill each vehicle go, when will the vehicle arrive and how many products are in eachvehicle.This article establishes two types of mixed-integer programming model based on thedifferent situation. The first is in the condition where the whole production can satisfy theneed of all customers. In this situation, we get three sub-type models: The basic model;the model that the vehicle can choose the way to customer through the warehouse or not;the model with the emergency transportation. The second is in the condition where someproducts can not satisfy the need of all customers. In this situation, we get another twosub-type models: the model that all customers are equal, so the assignment of the rareproducts is based on the principle which aims to achieve the goal of minimization oftransportation cost; the model that the position of each customer is different and thepriority one get satisfied first. In the end, we test all the models with the data from an oil company. We solves the model by using the mathematical programming solver CPLEX,we get two results. The first is optimized transportation cost getting close to the low boundof the total cost with the inventory increasing in the production line. The second is gettingtime when each vehicle should arrive at the factory, which will be a reference forenterprise to reduce the waiting time of the vehicle.There are some extensions in this Research. For example, the production-distributionnetwork only include one factory and one warehouse, which can be extend to complexnetwork including multiple factories and multiple warehouse. All the vehicles in themodel are the same, which can be extend to multi-vehicles. Moreover, the model is focuson making transportation plan after the production plan. It can be extended to synchronizeproduction and transportation plan.