Research On The Refined Product Multi-compartment Distribution Routing Optimization Problem

Refined product logistics exists in the whole petrochemical industry chain and connects the refinery,depot and petrol station to form the sales network.Currently.China's refined product consumption ranks third all over the world,which has become an important component of the economic development.The domestic refined product distribution mainly refers to the process of products delivered from depot to petrol stations and social customers.In recent years,the domestic petroleum companies have gradually adopted the initiative distribution mode to manage inventory and replenishment of affiliated petrol stations proactively.Therefore,how to realize the unified management of the distribution plan,reduce the transportation cost and improve the distribution efficiency have become the core issues of the refined product distribution management.The optimization of refined product distribution is a typical application of vehicle routing problem.In which,the delivery orders of gas stations need to be examined by the distribution center,collected in the depot,distributed by truck,inspection and unloading by gas stations sequentially,and involves a series of decisions such as the optimization of truck loading in depot,assignment of loading type in each trucks,and the optimization of the truck route.Beside,with different distribution strategies and application scenarios,the formulation criteria and methods of distribution schemes are diversified.Therefore,how to coordinate the optimization of product loading and driving route according to the existing resources,how to distribute the refined product from the depot to the gas station with higher efficiency on the premise of meeting different constraints and requirements,are the problems to be solved in this paper.Thus,this paper mainly conducts the following studies:1)Modeling and solving for the refined product multi-compartment and non-full load distribution routing problemConsidering the non-fully loaded in refined product distribution,to minimize the operating cost,two multi-compartment vehicle routing model are proposed according to the compartment loading type is fixed or flexible.For the multi-compartment vehicle routing model with fixed loaded product,with different transport strategies,two sub-models are derived.An ant colony and tabu search hybrid algorithm with different strategies is proposed to solve those two models.For the multi-compartment vehicle routing model with unfixed loaded product,introducing the time and space attributes of delivery orders,a two stage heuristic algorithm is designed for parallel execution of compartment assignment and route optimization.Besides,the applicability of different distribution strategies and the effectiveness of multi-compartment tuck are verified by above algorithms and simulation examples.2)Modeling and solving for the refined product multi-compartment multi-trip and full load distribution routing problemTo minimize total operating cost consists of transport cost and fixed cost,the replenishment time window is introduced and a multi-compartment routing model of multi-trip with heterogonous vehicles is developed.Then the problem is divided into two sub-problems based on the strategy of route-first and split-second.With an introduced concept of Adaptive Memory of Routes(AMR),a two-lever iterative algorithm is proposed.Though numerical experiments,the following rules are revealed:the time window with aggregating distribution can get highest time utilization,but with lower turnover time and more truck number;in addition,an effective traffic control and the vehicle configuration with flexible loading type can not only reduce the delivery cost but also improve the vehicle utilization.3)Modeling and solving for the refined product multi-compartment multi-trip distribution routing problem under bi-level decisionA bi-level programming model for refined product distribution is developed,in which the upper-level model is aim at the maximum profit of the transport company,the lower-level aims to maximize the satisfaction of gas stations.Compared with the previous two problems,this model starts from the perspective of double-layer decision and takes distribution volume and route as decision variables,which complying strictly with the capacity of oil tank in each station,the maximum unloading capacity of depot,the vehicle capacity and the load rate requirements.Based on the idea of interactive iterations,a two-lever interactive algorithm with a lower feedback strategy,genetic operator,sequential insertion operator,etc.is designed.Finally,with further experiment,the following rules is verified:The bi-level model is particularly suitable for the scene of transportation ability shortage;Increasing the unloading platform appropriately and optimizing the arrival time of trucks can effectively reduce the queue time;Beside,it is necessary to adjust truck configuration in time according to the sales situation to ensure the efficiency of using vehicle and the feasibility of the distribution plan.The research of this paper can further expand the theory and method of refined product multi-compartment distribution optimization,and provide new solution ideas.In addition,it also provide decision support for refined product sales companies and transportation companies.