Multimodal transport method and system in extremely cold environment

Abstract

The invention discloses a multimodal transport method and system in an extremely cold environment. The method comprises the following steps: constructing a multimodal transport low-temperature adaptive transport capacity grid model based on regional nodes and road sections; determining a conditional hypothesis of the model; constructing an optimization model taking the minimum total cost as a target; the total cost comprises basic transportation cost, transfer cost, loss / goods damage cost, energy consumption / refrigeration cost, equipment failure and maintenance cost and time window punishment and retention cost; establishing a mixed integer programming model, and determining constraint conditions; a hybrid particle swarm-immune dual optimization method is adopted to complete the above model solution iteration output optimal combined transport scheme; particles are constructed through three-section hybrid coding, and a double-optimization mechanism is established based on particle swarm local optimization continuous quantity and immune genetic global exploration discrete decision. And efficient, continuous and economical operation of multimodal transport in the extremely cold environment is guaranteed.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of extreme cold intermodal transportation, and more particularly relates to a multimodal transportation method and system in an extreme cold environment. BACKGROUND

[0002] Multimodal transportation is a mode of transportation that combines two or more modes of transportation, such as highways, railways, waterways, and aviation, to achieve integrated organization and operation of the entire transportation process, helping goods to arrive at the destination efficiently, continuously, and safely. It is an important mode of transportation in the modern logistics system that improves transportation efficiency. In recent years, with the increasing development of resources in high-latitude regions and cold regions, the demand for low-temperature transportation in the cold-chain logistics industry has continued to grow. In addition, international polar routes are gradually being developed and utilized, and the demand for multimodal transportation in extreme cold environments is increasing, becoming a new direction for development in the logistics field.

[0003] However, the current mainstream multimodal transportation technology and devices are mostly designed and operated based on temperate, subtropical, and other warm climate conditions. If they are directly applied to extreme cold environments (environmental temperatures are often between -30°C and -50°C or lower), they will face a series of special technical problems and challenges, severely restricting transportation efficiency.

[0004] First, the low-temperature reliability of devices and key components is insufficient. In extreme cold conditions, the lubrication performance of key components such as mechanical seals, bearings, hydraulic elements, and transmission chains is significantly degraded, the physical properties of materials are significantly reduced, and the brittleness is significantly increased. This can easily cause faults such as jamming, deformation, and even breakage, leading to a significant increase in equipment failure rate and maintenance frequency, directly reducing the continuous operation capability and overall availability of the multimodal transportation system.

[0005] Second, the power and temperature control support capability is degraded, making it difficult to maintain transportation continuity and goods quality. Low-temperature environments can cause battery storage capacity to decrease, electrical components to react slowly, and internal combustion engine power output to decline. Additionally, they can cause additional load on the temperature control system inside the container or cargo hold. For temperature-sensitive goods such as cold-chain food and pharmaceutical products, the existing technology lacks sufficient partition temperature control precision and real-time monitoring capability, making it impossible to ensure temperature stability during the entire transportation process, significantly increasing the risk of goods damage and enterprise operating costs.

[0006] Third, the adaptability of loading and unloading docking and transportation organization in extreme environments is insufficient. Ice, snow, and frost can adhere to the surfaces of transportation tool interfaces, quick docking mechanisms, and loading and unloading equipment, damaging their sealing performance and precision, and significantly reducing the efficiency of transfer and loading and unloading operations. At the same time, extreme cold weather can limit the capacity of infrastructure and delay the response speed of on-site maintenance. Traditional multimodal transportation organization modes and scheduling mechanisms lack the ability to flexibly respond to sudden weather and environmental constraints, making it difficult to ensure the stability of transportation time efficiency and emergency response efficiency.

[0007] The three types of problems are superimposed on each other, resulting in problems such as instability of operation, high cargo damage rate, large fluctuation of transportation time limit, and rising operation and maintenance cost in multimodal transport under extremely cold conditions, so it is urgent to propose targeted technical improvement schemes for low-temperature adaptability improvement, power and temperature control guarantee strengthening, and loading and unloading organization adaptability optimization. SUMMARY

[0008] The present application aims to solve the problems of insufficient low-temperature reliability of devices, degradation of power and temperature control guarantee, and poor adaptability of loading and unloading organization in multimodal transport under extremely cold environment. By constructing a low-temperature adaptation transport capacity grid model, an optimization model covering multiple types of costs and constraints is established, a hybrid particle swarm- immune double optimization algorithm is used for solution, a transport organization scheme suitable for extremely cold scenes is formed, the stability of equipment operation, the accuracy of cargo temperature control and the flexibility of transportation scheduling are improved, the cargo damage rate and the operation and maintenance cost are reduced, and the efficient, continuous and economic operation of multimodal transport under extremely cold environment is guaranteed.

[0009] In view of the above defects or improvement needs of the prior art, as a first aspect of the present application, the present application provides a multimodal transport method under extremely cold environment, comprising: S1. Constructing a multimodal transport low-temperature adaptation transport capacity grid model based on regional nodes and road sections; S2. Determining the conditional assumptions of the model; S3. Including decision variables such as road section selection, transfer selection and heating / energy scheduling into planning, and constructing an optimization model with the goal of minimizing total cost; the total cost includes basic transportation cost transfer cost loss / cargo damage cost energy consumption / refrigeration cost equipment failure and maintenance cost and time window penalty and detention cost ; S4. Combining the above costs to establish a mixed integer programming model, and determining the constraint conditions including path connectivity and flow conservation, temperature constraint, energy / endurance constraint, docking / node capacity constraint, time window constraint, and binary and physical variable domain constraint;

[0010] Further, the grid model in S1 has not less than 1 city / hub node as a vertex, not less than 1 connected road section as an edge, and the set of transportation modes on the edge is Each node is equipped with low-temperature adaptation device and interface unit, and its availability and energy supplement ability in extremely cold conditions are marked; The city / hub node includes a port, a railway marshalling station, a highway transfer field, and a polar scientific research station; the connected path section supports not less than one transportation mode.

[0011] Further, the condition assumption in S2 is that the goods are in the form of containers, the goods in the containers are not split for transportation, the low-temperature adaptation device and energy supplement facility marked in the region are in a usable state when planning, the transportation tool runs on the path section in accordance with the established driving rules, and each transfer node can be loaded and unloaded within the predetermined time window; all thermal / energy model parameters are estimated by field test or historical data and substituted into the model.

[0012] Further, the loss / cargo damage cost in S3 is: To describe the cargo freeze damage / quality deterioration caused by extremely cold, a Weibull three-parameter distribution is introduced to describe the loss rate over time : the Weibull three-parameter distribution form gives the failure probability density or cumulative distribution, and the loss rate is determined by the following formula: , where, is the scale parameter, is the shape parameter, is the location parameter; Based on the distribution, the cargo damage cost is: , where, is the number of containers used, is the value of unit box goods, is the total time of the goods in transit; if the cargo hold temperature control can be kept within the allowable range, then should be represented by the temperature correction function , is the effective temperature.

[0013] Further, the energy consumption / refrigeration cost in S3 is: The refrigeration and heating energy consumption caused by the cold chain and low-temperature adaptation mainly comes from the box heat loss, door opening / ventilation loss and equipment heating; the instantaneous heat loss of a certain partition of the cargo hold can be represented as: , where, represents the heat flow through the th partition; is the total heat transfer coefficient; For the first The heat transfer area of ​​each zone; The external ambient temperature sequence; For the first Temperature inside the structure; For the first Additional heat flow to each partition due to openings; Cooling / heating power requirements are ,in The performance coefficient is used to calculate the energy cost. , in, Indicates the first One mode of transportation; Price per unit of energy Including battery power and electric heating, This includes fuel heating and engine wear.

[0014] Furthermore, the time window penalty and dwell cost in S3 Specifically: , in, This is an indicator function; the function value is 1 when the condition inside the curly braces is met. The actual arrival station Time; For the site The latest allowed arrival time; arriving after this time is considered late. For the site The cost of penalties for being late; For the site The earliest allowed arrival time; anything earlier than this time is considered too early and needs to be waited for or stored. For the site The cost of early arrival, storage, or waiting.

[0015] Furthermore, the three-segment hybrid encoding in S5 includes: a path-mode binary segment, a relay / module activation segment, and a continuous segment; The path-mode binary segment is configured for each candidate path segment. With each method Use 0 / 1 encoding, or adopt a simplified encoding of "route node sequence + mode sequence"; The relay / module activation segment is represented by 0 / 1 encoding; The continuous range refers to the heating power. ,speed Real-value encoding is used.

[0016] Further, the fitness function in the S5 mixed particle swarm- immune dual optimization method is: , wherein, is the fitness; is the total cost.

[0017] As a second aspect of the present application, the present application provides a multimodal transport system in extremely cold environment, comprising: A low-temperature adaptive transport capacity grid construction unit is configured to construct a multimodal transport low-temperature adaptive transport capacity grid model based on regional nodes and road sections; A model condition assumption unit is configured to determine the condition assumption of the model; A total cost model unit is configured to incorporate road section selection, transfer selection and heating / energy scheduling decision variables into the planning, and construct an optimization model with the minimum total cost as the target; the total cost includes basic transportation cost transfer cost loss / damage cost energy consumption / refrigeration cost equipment failure and maintenance cost and time window penalty and detention cost. A mixed integer programming model unit is configured to combine the above costs, establish a mixed integer programming model, and determine the constraint conditions including path connectivity and flow conservation, temperature constraints, energy / endurance constraints, docking / node capacity constraints, time window constraints, and binary and physical variable domain constraints; A model solving unit is configured to complete the above model solving iteration to output the optimal transport scheme by using the mixed particle swarm- immune dual optimization method; wherein the particle is constructed by using a three-section mixed coding, and the dual optimization mechanism is established based on the local optimization of continuous variables of the particle swarm and the global exploration of discrete decisions of the immune genetic algorithm.

[0018] As a third aspect of the present application, the present application provides a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor to perform any step of the multimodal transport method in extremely cold environment.

[0019] Overall, compared with the prior art, the above technical solutions conceived by the present application can achieve the following beneficial effects: 1. The method for multimodal transport in extreme cold environment of the present application, by constructing a multimodal transport low-temperature adaptation transport grid model based on regional nodes and road sections, taking hub nodes such as cities, ports and railway marshalling stations as vertices, and connecting road sections supporting multiple transportation modes as edges, marking the availability and energy supplementing capacity of low-temperature adaptation devices (constant temperature cargo hold, heating docking device, etc.) of each node, and at the same time, making clear the model assumptions such as transportation of goods in units of containers, availability of node facility planning, providing a basic framework for transportation organization in extreme cold environment. This technical feature makes the coordinated work of transportation path and intermodal transport device can be accurately simulated, effectively solves the problem of disconnection between planning and reality caused by not considering the extreme cold adaptation in traditional intermodal transport model, and improves the adaptability and feasibility of transportation scheme in extreme low temperature scene.

[0020] 2. The method for multimodal transport in extreme cold environment of the present application, by selecting road sections, transfer selection, heating and energy scheduling as decision variables, constructing an optimization model with the minimum total cost as the objective, covering six types of costs such as basic transportation, transfer, cargo damage, energy consumption, equipment maintenance and time penalty, and at the same time, establishing a mixed integer programming model including path connectivity, temperature constraints (ensuring that the temperature of goods meets the category interval), energy endurance (matching battery and fuel capacity), etc., to realize the coordinated control of multi-dimensional cost and constraint. This technical feature solves the problems of high equipment failure rate, high cargo damage risk and uncontrollable energy consumption in extreme cold environment, which can reduce the frequency of equipment maintenance and cargo damage rate, reduce unnecessary energy consumption, and ensure the economy of transportation and the quality of goods.

[0021] 3. The method for multimodal transport in extreme cold environment of the present application, by using a hybrid particle swarm- immune double optimization algorithm to solve the optimization model, constructing particles with a hybrid coding of "path-mode binary segment + transfer / module activation segment + continuous real value segment", combining particle swarm algorithm to locally optimize continuous variables such as heating power and transportation speed, relying on immune genetic algorithm to conduct global exploration on discrete variables such as path selection and low-temperature adaptation module activation, and at the same time, evaluating the diversity of antibodies by Hamming distance and selecting excellent solutions to store in the memory bank. This technical feature efficiently solves the complex optimization problem of discrete decision and continuous scheduling interlaced in extreme cold intermodal transport, improves the solving efficiency and optimality of the solution compared with traditional algorithms, and can quickly output the transportation path, mode and parameter scheduling scheme adapted to extreme cold environment, ensuring the stability and timeliness of the whole transportation process. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 The flow chart of the method for multimodal transport in extreme cold environment of the embodiment of the present application; Figure 2 The schematic diagram of the multimodal transport low-temperature adaptation transport grid of the embodiment of the present application; Figure 3 The schematic diagram of the hybrid coding of the embodiment of the present application; Figure 4 This is a flowchart of the hybrid PSO-Immune algorithm according to an embodiment of the present invention; Figure 5 This is a system unit diagram of an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0024] Example 1 Please refer to Figure 1 This embodiment 1 provides a multimodal transport method for extremely cold environments, including: S1. Construct a multimodal transport low-temperature adaptive transport capacity grid model based on regional nodes and road segments; S2. Determine the conditional assumptions of the model; S3. Incorporate decision variables such as route selection, transfer selection, and heating / energy scheduling into the planning process to construct an optimization model with the objective of minimizing total cost; the total cost includes basic transportation costs. Transit costs Loss / damage costs Energy consumption / cooling costs Equipment failure and maintenance costs And time window penalties and retention costs Inside; S4. Combine the above costs to establish a mixed integer programming model, and determine the constraints including path connectivity and flow conservation, temperature constraints, energy / endurance constraints, docking / node capability constraints, time window constraints, and binary and physical variable domain constraints. S5. The above model is solved iteratively using a hybrid particle swarm optimization-immune dual optimization method to output the optimal transport scheme; the particles are constructed using a three-segment hybrid encoding method, and a dual optimization mechanism is established based on the continuous quantity of local optimization of particle swarm and the discrete decision of global exploration of immune genetics.

[0025] This embodiment 1 further elaborates on the above steps.

[0026] (1) Construction of low-temperature adaptive transportation capacity grid Please refer to Figure 2 This embodiment 1 constructs a multimodal transport low-temperature adaptive transport capacity grid model based on regional nodes and road segments. The model uses several city / hub nodes. (Including ports, railway marshalling yards, highway transfer stations, polar research stations, etc.) as vertices, with several connecting road sections (A road segment can support one or more modes of transportation: highway, railway, waterway, air, etc.) is an edge, and the set of transportation modes on the edge is . Each node is equipped with a cryogenic adaptation device and interface unit (temperature-controlled cargo compartment, heated docking station, energy replenishment station, etc.), and its availability and energy replenishment capabilities under extremely cold conditions are marked.

[0027] Based on the above grid, by selecting nodes, road segments and corresponding transportation modes, the collaborative work of cargo transportation routes and multimodal transport devices / systems in extremely cold environments under real-world conditions is simulated, and the organization mode selection and parameter scheduling are carried out with cost / risk / time as optimization objectives.

[0028] (2) Model condition assumptions To facilitate modeling and solving, the following assumptions are made, and these assumptions can be relaxed in specific implementations: Goods are transported in containers, and the goods inside the containers are not split up for transport (as assumed to be as a whole). The cryogenic adaptation devices and energy replenishment facilities marked in the area are in an available state during the planning stage (if they are not available, they are set to be inaccessible or have limited capacity in the model). The vehicles operate on the road segment according to the established driving rules, and each transfer node can be loaded and unloaded within the predetermined time window (if severely affected by ice conditions, this is reflected by the node accessibility parameter). All thermal / energy model parameters (such as the heat transfer coefficient of the vehicle compartment, PCM capacity, battery low-temperature degradation coefficient, and component failure distribution parameters) can be estimated through field tests or historical data and substituted into the model. The definitions and explanations of the decision variables and symbols in the model are further refined. The decision variables and symbols are as follows: Represents a set of city / hub nodes, indexed ; This represents a set of road segments, denoted as . ; Represents a set of transportation modes, index (1: Highway, 2: Railway, 3: Waterway, 4: Aviation, etc.)

[0029] Define decision variables; Indicates if in the road section The above adopts transportation methods The value is 1 if the route is selected, and 0 otherwise (Road Segment - Mode Selection). Indicates if at node Occurred by method Turn to The transit / docking operation is 1 (transit selection); This represents the number of containers used (which can be expanded to a vector of container counts for each route). Indicates the first Heating power used for critical components / cargo compartments at all times (can be used for decision-making during scheduling periods); Indicates road segment above method The operating speed (a controllable continuous variable); Define parameters and state variables; Indicates road segment The distance; Indicates road segment The above method The delivery time; Indicates unit container, unit distance in mode The basic unit price for transportation. Indicates at node transit from method arrive transit time; Indicates at node The unit transit cost for this transit; This indicates the volume or mass of a single container of goods, used for heat load calculations. Represents the external environment temperature sequence; Indicates the category of goods Permissible temperature range .

[0030] Define energy / equipment parameters; battery rated capacity At temperature Effective capacity correction factor The heat transfer coefficient of the constant temperature chamber PCM parameters , wait.

[0031] Define failure / loss parameters: critical component lifespan / failure distribution using Weibull's three parameters. express.

[0032] (3) Total Cost Model By incorporating decision variables such as route selection, transfer selection, and heating / energy scheduling into the planning process, an optimization model is constructed with the objective of minimizing total cost. The total cost consists of the following components, with their mathematical expressions provided.

[0033] Basic transportation costs When the road section The above method At that time, the basic transportation cost is related to the distance and the selected speed (speed affects energy consumption but also time): , in, This refers to the total number of containers transported or the number of containers carried on the corresponding route. Transit costs If at node Occurred by method Turn to For transit, transit costs and time must be included: , And accompanied by time increments .

[0034] Loss / damage cost (Weibull three-parameter distribution is used to describe cargo spoilage / freezing damage); To describe cargo freezing damage / quality deterioration caused by extreme cold, a Weibull three-parameter distribution is introduced to characterize the effects of time. Loss rate (or failure probability): The Weibull three-parameter distribution form gives the failure probability density or cumulative distribution, and the loss rate is determined by the following formula: , in, For scale parameters, For shape parameters, For position parameters; Based on this distribution, the cost of cargo damage is: , in, The number of containers used. The value of a unit box of goods. This refers to the total transit time of the goods; if the cargo hold temperature is controlled... If it can be kept within the allowable range, then Temperature correction function should be used. express, The effective temperature (the lower the temperature or the greater the temperature fluctuation, the higher the probability of loss).

[0035] Energy consumption / cooling cost The energy consumption for cooling and heating caused by cold chain and low-temperature adaptation mainly comes from heat loss of the cabinet, loss of air exchange due to door opening / exchange, and equipment heating (component insulation). a certain section of the cargo hold The instantaneous heat loss can be expressed as: , in, Indicates passing through the first Heat flow of each partition; The overall heat transfer coefficient; For the first The heat transfer area of ​​each zone; The external ambient temperature sequence; For the first Temperature inside the structure; For the first Additional heat flow to each zone due to openings (such as doors, windows, vents, hatches, etc.); Cooling / heating power requirements are ,in The performance coefficient is used to calculate the energy cost. , in, Indicates the first One mode of transportation; Price per unit of energy Including battery power and electric heating, This includes fuel heating and engine wear.

[0036] Equipment failure and maintenance costs (Based on Weibull model to describe component low-temperature failure rate); the failure probability of critical components (such as hydraulic pumps, bearings, and seals) can be represented by the Weibull model, which varies with operating time and low-temperature acceleration factor. Changes. Assume the expected equipment maintenance cost allocated to a single container is: , in, The probability of a component failing during a mission is calculated as follows: , in, The probability of a component failing within a given operating time; The mission runtime (the time the equipment operates continuously in extremely cold environments). These are the dimensional parameters (lifetime characteristic values) of the component. The low-temperature acceleration factor is used to represent the accelerated loss of lifespan due to extreme cold (the lower the temperature, the shorter the lifespan). For shape parameters (reflecting failure modes, such as early failure, random failure, or wear failure). This is a location parameter (the earliest possible time of failure). Time window penalty and retention cost If a delivery time window exists (e.g., arrival must not be later than...), or not earlier than Then, the cost of detention / penalty is introduced: , in, This is an indicator function; the function value is 1 when the condition inside the curly braces is met. The actual arrival station Time; For the site The latest allowed arrival time (upper limit of the time window); exceeding this time is considered late. For the site The costs of penalties for lateness (such as penalties for each delay, deductions in service quality points, and costs of breach of contract). For the site The earliest allowed arrival time (lower bound of the time window); anything earlier than this time is considered too early and needs to be waited for or stored. For the site Early arrival storage or waiting costs (e.g., costs incurred due to refrigeration of goods or waiting time for vehicles).

[0037] (4) Mixed Integer Programming Model Combining the above costs, we establish the following mixed integer programming (MINLP) model (which can be viewed as solving "optimal multimodal transport unit / system organization and parameter scheduling under extremely cold conditions"); its optimization objective is to minimize the total cost, as follows: , At the same time, the following constraints are established: Path connectivity and flow conservation (ensuring flow from the originating node) to the destination node There is a feasible path: , in, This represents an intermediate node.

[0038] Temperature constraints (the temperature of goods must be maintained during each leg of the journey and during transshipment): , in, Indicates the category of goods.

[0039] Energy / range constraints (battery and fuel capacity limitations): , in, Indicates at time Power consumption (kW); Indicates the time step, used for integral approximation; This indicates the total available battery energy capacity of the system (kWh). Indicates at time Fuel consumption power (kW); This represents the total fuel energy available for the transportation mission (in kWh or volumetric equivalent).

[0040] Docking / node capacity constraints (node ​​transfer throughput, maximum power of heating facilities, etc.): , in, Indicates at node Whether it occurs due to the mode of transportation Steering method The transfer (a binary variable, 1 indicates that it has occurred, and 0 indicates that it has not occurred); This indicates the number of cargo units transshipped (usually measured in containers). Represents a node Maximum transshipment capacity (number of containers that can be transshipped per unit of time). Indicates the node at time [time]. Total power requirement (kW) for heating or cooling facilities; This indicates the maximum heating / cooling power (kW) that the node can provide.

[0041] Time window constraint (arrival time meets customer time limit): , in, Indicates the arrival time of the vehicle or goods; Indicates the delivery deadline requested by the customer; Indicates the allowed delay buffer time; This indicates the earliest time when the goods can be received (to prevent premature arrival and the resulting backlog or storage costs).

[0042] Binary and physical variable domain constraints: , in Indicates whether to select from the node To the node The The path (binary variable) for each mode of transportation; Indicates whether it is in the node A mode of transport conversion occurs (binary variable); Indicates at time The heating power input (kW) must be non-negative; Indicates the mode of transport In the path The operating speed is within the range of ; Indicates the mode of transport Minimum running speed on path (i,j); Indicates the mode of transport In the path The maximum operating speed.

[0043] This model is a mixed integer nonlinear / nonconvex problem (due to thermal-energy coupling and the Weibull term), which can be solved by approximation / linearization or metaheuristic algorithms.

[0044] (5) Model Solving Please refer to Figure 4 To efficiently solve the complex problem involving discrete path selection and continuous scheduling decisions, a Hybrid PSO-Immune algorithm is proposed. This algorithm borrows from and improves the immune encoding and evolution mechanism in the template, while employing local optimization of particle swarm for continuous control variables (such as heating power).

[0045] Decoding / antibody (particle) encoding method; please refer to Figure 3 In a preferred embodiment, the hybrid encoding method designs the antibody (particle) as a string containing a path-mode binary segment, a transit / module activation segment, and a continuous segment. The specific encoding rules are as follows: The path-mode binary segment uses compressed encoding of "route node sequence + mode sequence". The route node sequence marks candidate road segments with 0 / 1. The selection status is 1, which indicates that the road segment is selected, and 0 indicates that it is not selected. The continuity of the path is ensured through the node connection relationship. The transportation mode sequence uses 1 / 2 / 3 / 4 to correspond to road / rail / waterway / air transportation respectively, and corresponds one-to-one with the node sequence to form a chain structure of "node A-1-node B-3-node C", which accurately describes the combination scheme of transportation path and mode.

[0046] The transfer / module activation segment is a 0 / 1 binary sequence, with each bit corresponding to the activation status of a specific functional module: the first bit indicates whether the node is using the cryogenic adaptation module (1 for enabled / 0 for disabled), the second bit indicates whether the heating docking device is activated (1 for activated / 0 for disabled), and the third bit indicates whether phase change material (PCM) thermal storage is used (1 for used / 0 for not used). Subsequent bits can be extended to other extreme cold adaptation functions, and the digital expression of multi-module collaborative control is realized through binary combination.

[0047] The continuous measurement segment uses real-value encoding to directly record the heating power. Speed ​​of travel on road sections The specific values ​​of continuous variables are retained to two decimal places; real-valued data are directly used as the optimization object of the particle swarm algorithm, and the parameters are finely adapted through local optimization.

[0048] The three code segments are concatenated to form a complete antibody (particle) string, such as "101101-123-110-25.68-90.50". The first 6 bits are the binary path selection, the middle "123" is the transport mode sequence, "110" indicates that the cryogenic module is enabled and the heating docking is not used, and the last two real values ​​correspond to the heating power and the travel speed, respectively, realizing a unified encoding expression of discrete decision and continuous parameters.

[0049] The fitness function in this embodiment 1 uses the total cost As the core evaluation criterion, it is constructed using a reciprocal form, with the specific expression as follows: , This function enables a quantitative evaluation of the quality of antibodies (particles), adapting to the breeding and selection strategies of the immune algorithm.

[0050] Of which, total cost The study covers six core cost categories for multimodal transport in extremely cold environments: basic transportation costs (including route fees for different modes of transport), transshipment operation costs (node ​​loading / unloading and connection expenses), cargo loss costs (calculated based on a loss function of temperature deviation and time), energy consumption costs (total expenditure on electricity and fuel), equipment maintenance costs (the sum of the product of component failure probability and repair costs), and time penalty costs (additional expenditures for exceeding planned delivery times). By summing these costs, a comprehensive indicator reflecting the economic efficiency and feasibility of the solution is formed.

[0051] The inverse form of the fitness function keeps the evaluation logic consistent with the optimization objective: when The smaller the size (the lower the cost of the solution). The closer the value is to 1 (the higher the fitness); when Increase (the cost of the solution rises). The value decreases accordingly, which aligns with the decision criterion that "lower-cost solutions are better." This design preserves the monotonic mapping relationship between cost and fitness, while also facilitating fitness-based antibody selection in the immune algorithm. During the breeding phase, antibodies with high fitness are more likely to be selected for crossover and mutation, thus driving the population towards a lower-cost solution. During clonal selection, excellent antibodies are selected and stored in the memory bank through a fitness threshold, achieving the retention and iterative optimization of high-quality solutions.

[0052] At the same time, this function also takes into account the implicit evaluation of constraints: for schemes that violate constraints such as temperature range, total energy, and path connectivity, its The fitness value will be significantly amplified through the penalty mechanism (such as introducing a penalty for constraint violation), resulting in a sharp drop in the fitness value, and thus being naturally eliminated in the population evolution, indirectly achieving the screening of the feasibility of the solution.

[0053] Further, in a preferred embodiment, during the initialization phase, an antibody / particle population of size F is constructed, and an initial solution containing a complete coding structure is generated using a template. The path-mode segment randomly generates feasible paths based on the transportation network topology, ensuring connectivity from the origin to the destination, and the transportation mode is allocated according to the probability of road segment adaptability; the transfer / module activation segment uses random 0 / 1 assignment to achieve the initial combination of functional modules; the continuous quantity segment is initialized using a uniform distribution within an engineering experience interval, and the driving speed is randomly generated according to the transportation mode threshold. The initial state of the memory bank can optionally store 2-3 sets of heuristic solutions, including those based on... The algorithm's shortest time path (prioritizing air travel + high-grade highways) and lowest energy consumption path based on the energy consumption model (prioritizing rail travel + waterways) provide high-quality initial guidance for population evolution.

[0054] In a preferred embodiment, to avoid the algorithm getting trapped in local optima and maintain the population's evolutionary vitality, antibody diversity evaluation and superior individual selection need to be completed through a dual-dimensional approach. First, for the antibody's path-mode binary segment and relay / module activation discrete segment, Hamming distance is used to quantify the similarity between individuals: the number of bits with different values ​​in corresponding positions of the two encoding segments is counted, and the ratio of this ratio to the total encoding length is used as a similarity index. Simultaneously, a similarity threshold is set. .

[0055] Antibody concentration is calculated based on similarity, i.e., the concentration of antibodies in the population that have a similarity to the target antibody exceeds a certain threshold. The proportion of individuals with high fitness to the total population size is considered. Higher concentrations indicate a denser "similar cluster" to which the antibody belongs, requiring a reduction in its reproductive priority to maintain diversity. Subsequently, a desired reproductive probability model is constructed to balance the need for "high-fitness preferential reproduction" with "low concentration to avoid homogenization".

[0056] The selection phase employs a roulette wheel selection method, allocating the probability of individuals being selected based on their expected reproductive probability. This ensures that superior antibodies (high fitness, low concentration) receive more reproductive opportunities while preserving evolutionary space for low-concentration potential antibodies. Simultaneously, individuals with high fitness rankings and concentrations below a certain threshold are selected. Individuals with good fitness are defined as excellent individuals and are directly stored in the memory bank. When the memory bank reaches its maximum capacity, individuals with the lowest fitness are removed, thereby achieving dynamic updates to the memory bank and continuous accumulation of high-quality solutions.

[0057] In a preferred embodiment, a dual optimization strategy combining particle swarm optimization and immune genetics is employed to achieve synergy between fine-tuning of continuous parameters and global exploration of discrete decision-making. Particle swarm optimization focuses on optimizing real-value segments, performing PSO updates on heating power and road segment speed; individual optimal (pbest) and swarm optimal (gbest) methods guide the continuous parameters towards a better solution.

[0058] Global exploration of immune genetics involves operations applied to discrete segments: the selection operator retains the top 50% of individuals by fitness; the path-mode segment uses two-point crossover, exchanging gene fragments in random intervals of the node sequence; the transit / module activation segment uses single-point crossover, exchanging binary sequences at random positions. Mutation operations adjust probabilities using an iterative decay strategy, randomly flipping 1-2 selection bits for path segments and randomly switching 1 activation state for module segments to avoid premature convergence of the population.

[0059] After the dual optimization results are merged, they are re-encoded and their fitness is calculated, updating the individual best (pbest) and the group best (gbest). The memory bank includes the newly generated top 3 excellent individuals in each generation. When duplicate solutions exist, those with higher fitness are retained. Through dynamic balancing of "local fine-tuning + global optimization", multi-dimensional optimization of the extreme cold intermodal transport scheme is achieved.

[0060] When the iteration upper limit is satisfied When the fitness converges, the optimal antibody in the output memory is selected as the optimal organization mode and system parameter configuration, including the path, mode selection, transfer node activation strategy, and continuous control quantities (heating power, temperature).

[0061] This embodiment 1 also provides a specific implementation step, as follows: Step 1: Regional Capability and Mode Prediction: Based on the low-temperature adaptability of each node in the country / region (whether there are constant temperature chambers, refueling stations, de-icing capabilities, etc.), the node capability level is given and the appropriate intermodal transport organization mode category is preliminarily determined (e.g., road-rail intermodal transport with node heating enabled; road-water intermodal transport requires the use of high-power heating docking at the port, etc.).

[0062] Step 2: Establish a basic low-temperature transportation capacity grid: Select several large and medium-sized cities / hubs in the study area as nodes, and construct a grid database of road segments, transportation modes and node capabilities (including distance, annual minimum temperature, node energy replenishment / de-icing capacity, throughput capacity, cost parameters, etc.).

[0063] Step 3: Construct a thermal-energy and loss parameter library: Calibrate the cargo compartment heat transfer coefficient UA, PCM parameters, component Weibull parameters, and battery performance derating function using experimental or empirical data. wait.

[0064] Step 4: Establish a joint optimization model: Establish the total cost target and constraints according to Example 1, and set the range of decision variables and time discretization (e.g., with hours as the step size).

[0065] Step 5: Set up and run the hybrid PSO-Immune algorithm to solve the problem: Set the population size F, memory capacity, crossover / mutation probability, PSO parameters (inertia weight, learning factor), and maximum number of iterations. Then, run the algorithm to obtain the optimal solution.

[0066] Step 6: Simulation verification and parameter fine-tuning: Perform thermal-energy dynamic simulation of the optimal solution in MATLAB / Simulink or Python environment to verify temperature constraints and energy consumption budget; adjust model parameters or algorithm weights according to simulation results and solve again.

[0067] Step 7: On-site implementation: Send the organization mode (path, mode, transit node and module activation command) output by the algorithm to the scheduling system, and execute the corresponding self-heating, lubrication preheating, de-icing and docking processes at the vehicle / ship and node ends.

[0068] Step 8: Online Re-optimization: During cargo transportation, the system receives real-time node availability and equipment status information, triggering online replanning (a fast version of the MPC + restart hybrid algorithm) to cope with sudden extreme cold events.

[0069] Example 2 Please refer to Figure 5 This embodiment 2 provides a multimodal transport system for extremely cold environments, including: Low-temperature adaptable transport capacity grid building unit is used to construct a multimodal transport low-temperature adaptable transport capacity grid model based on regional nodes and road segments; Model condition assumption unit, used to determine the condition assumptions of the model; The total cost model unit is used to incorporate decision variables such as route selection, transfer selection, and heating / energy scheduling into the planning process, constructing an optimization model with the objective of minimizing total cost; the total cost includes basic transportation costs. Transit costs Loss / damage costs Energy consumption / cooling costs Equipment failure and maintenance costs And time window penalties and retention costs Inside; The mixed-integer programming model unit is used to combine the above costs, establish a mixed-integer programming model, and determine the constraints including path connectivity and flow conservation, temperature constraints, energy / endurance constraints, docking / node capability constraints, time window constraints, and binary and physical variable domain constraints. The model solving unit is used to complete the above model solving iterative output of the optimal joint transport scheme using a hybrid particle swarm-immune dual optimization method; wherein the particles are constructed by a three-segment hybrid encoding, and a dual optimization mechanism is established based on the continuous quantity of local optimization of particle swarm and the discrete decision of global exploration of immune genetics.

[0070] Example 3 This embodiment 3 also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement any step of a multimodal transport method in an extremely cold environment.

[0071] The computer-readable storage medium may include various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0072] For a description of the computer-readable storage medium provided in this application, please refer to the above method embodiments; further details will not be repeated here.

[0073] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for multimodal transport in an extremely cold environment, characterized by The method comprises the following steps: S1. Constructing a multimodal transport low-temperature adaptation transport grid model based on regional nodes and road sections; S2. Determining the conditional assumptions of the model; S3. Incorporate route selection, transshipment selection, and heating / energy scheduling decision variables into the planning, and construct an optimization model with the objective of minimizing the total cost, including the base transportation cost , transshipment cost , loss / damage cost , energy consumption / refrigeration cost , equipment failure and maintenance cost , and time window penalty and detention cost ​ S4. Combining the above costs to establish a mixed integer programming model, and determining the constraint conditions including path connectivity and flow conservation, temperature constraints, energy / endurance constraints, docking / node capacity constraints, time window constraints, and binary and physical variable domain constraints; S5. Using a hybrid particle swarm-immune dual optimization method to complete the above model solving iteration to output the optimal intermodal transport scheme; wherein the particles are constructed in a three-segment hybrid coding, and a dual optimization mechanism is established based on the local optimization of continuous quantities of particle swarm and the global exploration of discrete decisions of immune genetic.

2. The method of multimodal transport in extreme cold environment according to claim 1, characterized in that, The S1 grid model has not less than 1 city / hub node as a vertex, and not less than 1 connected path segment as an edge as an edge, and the set of transportation modes on the edge is Each node is equipped with a low-temperature adaptation device and an interface unit, and is marked with its availability and energy supplement capacity under extremely cold conditions. The city / hub node includes a port, a railway marshalling station, a highway transfer field, and a polar scientific expedition station; and the connected road section supports not less than one transportation mode.

3. The method of multimodal transport in extreme cold environment according to claim 1, characterized in that, The conditional assumptions in S2 are as follows: the goods are in the form of containers, the goods in the containers are not split for transportation; the low-temperature adaptation devices and energy supplement facilities marked in the region are in a usable state during planning; the transportation tools run on the road sections in accordance with the established driving rules, and each transfer node can be loaded and unloaded within the predetermined time window; and all thermal / energy model parameters are estimated by field tests or historical data and substituted into the model.

4. The method of multimodal transport in polar conditions according to claim 1, characterized in that, The S3 loss / damage cost Specifically: To describe the freezing damage / quality deterioration of goods caused by extreme cold, a Weibull three-parameter distribution is introduced to characterize the rate of loss over time The Weibull three-parameter distribution form gives the failure probability density or cumulative distribution, and the rate of loss is determined by the following equation: ， wherein, is a scale parameter, is a shape parameter, is a position parameter; Based on the distribution, the cargo loss cost is: ， in, The number of containers used. The value of a unit box of goods. This refers to the total transit time of the goods; if the cargo hold temperature is controlled... If it can be kept within the allowable range, then Temperature correction function should be used. express, The effective temperature.

5. The method of multimodal transport in polar conditions according to claim 1, characterized in that, The energy consumption / refrigeration cost in S3 Specifically: The refrigeration and heating energy consumption caused by cold chain and low temperature adaptation mainly comes from the box heat loss, door opening / ventilation loss and equipment heating; the instantaneous heat loss of a certain partition of the cargo hold can be expressed as: ， wherein, represents the heat flow through the first partition; is the overall heat transfer coefficient; is the heat transfer area of the first partition; is the external ambient temperature sequence; is the temperature inside the first structure; is the additional heat flow due to the opening of the first partition; The refrigeration / heating power demand is wherein is the coefficient of performance, thus the energy cost: ， wherein, represents the transportation mode; is the unit energy price, includes battery power and electric heating, includes fuel heating and engine losses.

6. The method of multimodal transport in polar conditions according to claim 1, characterized in that, The time window penalty in S3 and the detention cost Specifically: ， where, is an indicator function, the function value is 1 if the condition in the braces is satisfied; is the actual arrival time at the station ; is the latest allowed arrival time at the station , exceeding which counts as late; is the late penalty cost at the station ; is the earliest allowed arrival time at the station , being earlier than which counts as early and requires waiting or storage; is the early arrival storage or waiting cost at the station .

7. The method of multimodal transport in polar conditions according to claim 1, characterized in that, The three-segment hybrid coding in S5 includes a path-mode binary segment, a transfer / module activation segment, and a continuous quantity segment; The path-way binary segment is configured for each candidate path segment with each way using 0 / 1 encoding, or using "route node sequence + way sequence" compact encoding; The transfer / module activation segment is coded by 0 / 1; The continuous quantity section is a heating power , speed Real value encoding is used.

8. The method of multimodal transport in polar conditions according to claim 1, characterized in that, The fitness function in the hybrid particle swarm-immune dual optimization method in S5 is: ， wherein, is the fitness; is the total cost.

9. A multimodal transport system in an extremely cold environment, characterized by The method comprises the following steps: A low-temperature adaptation transport grid construction unit is configured to construct a multimodal transport low-temperature adaptation transport grid model based on regional nodes and road sections; A model conditional assumption unit is configured to determine the conditional assumptions of the model; A total cost model unit is configured to incorporate the decision variables of route selection, transshipment selection, and heating / energy scheduling into the planning, and to construct an optimization model with the objective of minimizing the total cost; the total cost includes basic transportation cost , transshipment cost , loss / damage cost , energy consumption / refrigeration cost , equipment failure and maintenance cost , and time window penalty and detention cost ; A mixed integer programming model unit is configured to combine the above costs to establish a mixed integer programming model, and determine the constraint conditions including path connectivity and flow conservation, temperature constraints, energy / endurance constraints, docking / node capacity constraints, time window constraints, and binary and physical variable domain constraints; A model solving unit is configured to use a hybrid particle swarm-immune dual optimization method to complete the above model solving iteration to output the optimal intermodal transport scheme; wherein the particles are constructed in a three-segment hybrid coding, and a dual optimization mechanism is established based on the local optimization of continuous quantities of particle swarm and the global exploration of discrete decisions of immune genetic.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by a processor to perform the method of claim 1-8.

Images