CN-121981633-A - Hub branch road and land port system freight dispatching method based on intelligent prediction and branch pricing algorithm

CN121981633ACN 121981633 ACN121981633 ACN 121981633ACN-121981633-A

Abstract

The invention discloses a freight dispatching method of a hub branch road land port system based on intelligent prediction and branch pricing algorithm, which relates to the technical field of container shift transport network dispatching, and the method comprises the steps of carrying out short-term demand prediction by collecting container loading and unloading demand data of each port and combining with an intelligent prediction method machine learning model, constructing a mixed integer linear programming model based on a three-level port network structure of a hub port, a branch port and a land port, and controlling ship access port sequence, departure time, speed configuration and connection transport coordination through genetic algorithm, simulated annealing algorithm, branch pricing algorithm and a mixed strategy thereof, so as to realize a model target of minimizing total navigation cost and transport time; according to the invention, under a complex transportation network level, through deep fusion of machine learning prediction and hybrid optimization algorithm, the key problems of poor dynamic adaptability, prediction and optimization disjoint and the like in traditional shipping scheduling are solved, and the transportation efficiency is improved, and the energy and environmental performance are improved.

Inventors

CAO ZHICHAO
BAI RAN
ZHAO ZICAN
QIAN TAO
Gu Jinggang
LI MAN
WANG YUMING
XU XUNQIAN
SHI QUAN
SUN JINGXUAN
HOU MINGYAN

Assignees

南通大学

Dates

Publication Date: 20260505
Application Date: 20260123

Claims (8)

1. The hub branch road and land port system freight dispatching method based on the intelligent prediction and branch pricing algorithm is characterized by comprising the following steps of: S1, constructing a three-layer transportation network model comprising a hub port, a branch port and an abdominal port, and defining a port topological structure, a transportation relationship and a cargo flow direction; S2, constructing a port dynamic demand prediction model based on historical loading and unloading amount data, predicting the demand of each port container in a future period, generating a candidate scheduling scheme of a main route and a connection route according to the predicted demand, and establishing a comprehensive optimization model comprising a route time window, fuel consumption, loading and unloading operation time and cost; And S3, a method for combining a branch pricing algorithm and a heuristic algorithm of the design column generating framework is adopted, and the comprehensive optimization model is solved, so that an optimal route scheduling scheme which meets the transportation requirement and has the minimum total cost is obtained.
2. The method for dispatching freight in a hub branch road and land port system based on intelligent prediction and branch pricing algorithm according to claim 1, wherein in step S1, the three-layer transportation network model comprises a hub port set Ω, a branch port set H and a land port set C, and the ships K are classified into two types of main-line ships and branch-line ships; The main line ship starts from the terminal port, traverses the branch port with loading requirement on the way, finishes loading and unloading service, returns to the terminal port at the earliest time, and simultaneously, the branch line ship starts from the land port, and transports the goods at the land port to the branch port with loading and unloading requirement, and the transportation route is carried out simultaneously, but the branch line ship must arrive before the main line ship arrives at the branch line port.
3. The hub branch road and land port system freight scheduling method based on intelligent prediction and branch pricing algorithm according to claim 1, wherein in step S2, the port dynamic demand prediction model predicts by adopting a machine learning long-short-term memory network LSTM prediction model, and the LSTM prediction model processes input data including historical throughput, weather indexes and economic indexes through a gating mechanism.
4. The hub branch road and land port system shipment scheduling method based on intelligent prediction and branch pricing algorithm according to claim 2, wherein in step S2, the comprehensive optimization model comprises the following targets: Wherein, the Representing vessels as decision variables Whether or not to go from port Travel to port The integral objective function is formed by The connection is carried out, Indicating whether the goods originating point is connected and transported, if yes, 1, if no, 0, Representing the ship K from the port Travel to port Is used for the production of the high-density polyethylene, Representing the fixed cost of the cargo at the starting point of the cargo transport of the connection ship, the cost structure is divided into two main parts, namely the cost of the trunk line and the cost of the branch connection, including the operation fees of various ships used on the trunk line and the branch line and the loading and unloading fees of the container In the following In the process, the And (3) with Respectively represent that the ship is at the port Is used for the loading and unloading of the conveyor belt, Representing the cost per unit load/unload; Representing a ship Dynamic costs during voyage, wherein, Indicating the cost of fuel consumption, Representing the cost of vessel rental, relating to the type of vessel rented and the total length of the voyage, Representing environmental emission costs, by minimizing the weighted sum of these three components, the comprehensive optimization model achieves a balance between energy economy, transport efficiency, and environmental benefits, fixed costs being determined by Definition, each component has definite operational and economic significance, and the complete mathematical expression of the comprehensive optimization model is as follows ; Wherein, the A multiplicative amplification representing sea state and weather versus fuel consumption, The value range of (2) is 0.5-0.7, The range of values is 0.3-0.5, the accurate value depends on the dispatching scale, and if complex multi-port operation is involved, the parameter weight is close to 1.
5. The hub branch road and land port system shipment scheduling method based on intelligent prediction and branch pricing algorithm according to claim 4, wherein in step S2, the ship speed effect on fuel consumption satisfies a functional relationship: , wherein, Representing a base fuel consumption coefficient; the navigational speed power representing the control energy consumption curve reflects the sharp increase in fuel consumption at high speed voyages.
6. The hub branch road and land port system shipment scheduling method based on intelligent prediction and branch pricing algorithm according to claim 4, wherein in step S2, the comprehensive optimization model comprises: ensuring that each vessel starts from the terminal port and eventually returns to the same terminal port, Ensuring that any vessel entering a port is launched from that port; Ensuring that each point of origin of the good is assigned to exactly one branch port, Ensuring that the total amount of cargo from the cargo origin point to the port does not exceed the capacity of the ship; Ensuring a ship Branch port After the loading and unloading operation is completed, the residual capacity of the loading and unloading operation must meet the requirement of the next branch port Is not limited by the loading and unloading requirements of the vehicle; ensuring that the ships are transported to branch ports in sequence and returned to the terminal port before the latest allowable return time; Ensuring that the vessel performs branch port transportation in sequence and returns to the terminal port before the latest return time allowed; ensuring that the spur vessel transfers cargo from the origin point to the spur port before the arrival of the main vessel; Limiting ship Branch port Is located within the transit time window of the port.
7. The hub branch road and land harbour system shipment scheduling method based on intelligent prediction and branch pricing algorithm of claim 1, wherein in step S3: Generating an initial feasible route and speed distribution scheme by combining a heuristic algorithm and a simulated annealing algorithm, constructing a main problem which takes a ship type and a route column as decision objects, constructing a sub-problem by utilizing a shortest path strategy, continuously introducing a new route column for reducing cost by a column generation mechanism, and solving integer variables based on a set branching strategy until a global optimal scheduling scheme is obtained; The method comprises the steps of describing a feasible route set by a main problem by adopting a set coverage model, generating a column for reducing cost by adopting a shortest path pricing model by a sub-problem, triggering a branch mechanism to ensure the feasibility of integers when nodes are infeasible or variables collide, firstly solving an initial solution by using GA and SA algorithms, then solving a mixed branch pricing algorithm, and finally comparing and analyzing the three algorithms, a solver Gurobi and a combination thereof based on a numerical experiment to evaluate the performance difference of the three algorithms in terms of the quality and the computational efficiency of the solutions.
8. The hub branch road and land harbour system freight scheduling method based on intelligent prediction and branch pricing algorithms according to claim 1, wherein in step S3, the output results comprise a main route access sequence, a connection route access sequence, the speed and time of each route segment, the fuel consumption of each ship and the overall route cost under each algorithm and a final scheme visualization.

Description

Hub branch road and land port system freight dispatching method based on intelligent prediction and branch pricing algorithm Technical Field The invention relates to the technical field of container shift transport network dispatching, in particular to a hub branch road and land port system freight dispatching method based on intelligent prediction and branch pricing algorithm. Background With the continuous growth of global trade, container shipping networks are increasingly complex, and shipping scheduling at multiple ports becomes a key factor affecting transportation efficiency, operating costs, and quality of service. The traditional ship scheduling method mainly depends on a static network model and historical data assumption, and is difficult to effectively process dynamic factors such as port demand fluctuation, ship speed adjustment, fuel consumption change and weather influence, so that resource distribution is uneven, cost is high, and robustness of quasi-point control is insufficient. Therefore, research into a scheduling technology of dynamic prediction, network modeling and optimization algorithms has become an urgent need for improving the reliability and economy of maritime logistics. The multi-stage harbor shift ship scheduling problem is essentially a large-scale combinatorial optimization problem and has NP-hard characteristics. The existing research is mainly developed from three layers of network structure, demand prediction and algorithm case scale, but the method limitation still exists: (1) Network modeling is simplified-early studies such as Yue and Mangan (2024) employ a static reliability framework, focus network topology analysis, but do not incorporate dynamic scheduling constraints, resulting in a model with insufficient adaptability in practical operation. Xu et al (2024) describe global port associations by multi-network theory, providing structural insight, but lacking a combination with optimization algorithms, cannot address precise scheduling issues. (2) The prediction is separated from optimization, and most methods such as Tamburini et al (2023) rely on historical data to statically assume demand and cannot respond to market fluctuations. Wang et al (2024) developed a multivariate hybrid system for port throughput prediction, combining ARIMA and LSTM models to achieve mape=7.8% accuracy, but the prediction results were decoupled from the scheduling decisions and not used for precise optimization. (3) The algorithm efficiency and the scale are that an accurate algorithm (such as branch pricing B & P) can guarantee an optimal solution but the calculation cost is high, and a heuristic algorithm (such as a genetic algorithm GA and a simulated annealing SA) is high in calculation efficiency but easy to fall into local optimal. For example, zhao et al (2024) use GA to achieve about 5% cost savings, but assume that all ports are directly accessible, in the case of a small number of ports, and disjoint from the hub-spur hierarchy in real streams. Aiming at the defects, the invention researches the multi-stage port shift scheduling to realize the dynamic prediction of the port container and the algorithm optimization of the scheduling line under the level of the complex transportation network, and provides scientific decision-making technical support for shipping companies. Disclosure of Invention The invention aims to provide a hub branch road and land port system freight dispatching method based on intelligent prediction and branch pricing algorithm, so as to solve the problems in the prior art in the background art. In order to achieve the above purpose, the present invention provides the following technical solutions: The hub branch road and land port system freight dispatching method based on the intelligent prediction and branch pricing algorithm comprises the following steps: S1, constructing a three-layer transportation network model comprising a hub port, a branch port and an abdominal port, and defining a port topological structure, a transportation relationship and a cargo flow direction; S2, constructing a port dynamic demand prediction model based on historical loading and unloading amount data, predicting the demand of each port container in a future period, generating a candidate scheduling scheme of a main route and a connection route according to the predicted demand, and establishing a comprehensive optimization model comprising a route time window, fuel consumption, loading and unloading operation time and cost; And S3, a method for combining a branch pricing algorithm and a heuristic algorithm of the design column generating framework is adopted, and the comprehensive optimization model is solved, so that an optimal route scheduling scheme which meets the transportation requirement and has the minimum total cost is obtained. Preferably, in step S1, the three-layer transportation network model includes a hub port set Ω, a branch port set H, and a land port set