A warehouse logistics management method based on logistics big data

Abstract

The application belongs to the technical field of logistics management, and discloses a warehouse logistics management method based on logistics big data, which comprises the following steps: collecting logistics multi-source heterogeneous data; preprocessing the logistics multi-source heterogeneous data to obtain preprocessed heterogeneous data; performing mode recognition on the preprocessed heterogeneous data by using an improved automaton algorithm to obtain M kinds of logistics route modes; based on the M kinds of logistics route modes, an optimal logistics planning scheme is calculated by using an optimization algorithm, and the optimal logistics planning scheme is used to guide warehouse logistics management, so that the logistics response speed and service quality are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics management, and more particularly, to a warehouse logistics management method based on logistics big data. BACKGROUND

[0002] The existing logistics management technology has exposed many problems in the face of increasingly complex supply chain environment and massive data. First, the traditional method is difficult to effectively process and integrate heterogeneous data from different sources, resulting in that the data value cannot be fully mined and the decision basis is not comprehensive enough. Second, the existing route planning algorithm often only considers static factors and cannot respond to dynamic factors such as traffic conditions and weather changes, resulting in low actual transportation efficiency. Third, the traditional inventory management strategy lacks flexibility and is difficult to respond to rapid changes in market demand, often resulting in inventory overstock or shortage. In addition, the existing logistics optimization method usually only focuses on a single target such as cost minimization, while ignoring multi-dimensional indicators such as timeliness and service quality, making it difficult to meet the comprehensive needs of modern logistics. In actual operation, logistics enterprises often face problems such as information silos, unreasonable resource scheduling, poor emergency response capability, and other problems, which seriously affect the overall operational efficiency. More importantly, it is difficult to achieve global optimization of the entire supply chain, resulting in poor coordination between supply chain links and weak risk resistance. These problems are particularly prominent in special circumstances such as e-commerce promotion and public emergencies.

[0003] In view of this, the present application provides a warehouse logistics management method based on logistics big data to solve the above problems. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, in order to achieve the above-mentioned purpose, the present application provides the following technical scheme: a warehouse logistics management method based on logistics big data, characterized in that it comprises: S1, collecting logistics multi-source heterogeneous data;

[0005] S2, preprocessing the logistics multi-source heterogeneous data to obtain preprocessed heterogeneous data;

[0006] S3, using an improved automaton algorithm to perform pattern recognition on the preprocessed heterogeneous data to obtain M kinds of logistics route patterns;

[0007] S4, based on the M kinds of logistics route patterns, an optimal logistics planning scheme is calculated and obtained, and the optimal logistics planning scheme is used to guide warehouse logistics management.

[0008] Further, the logistics multi-source heterogeneous data includes logistics transportation data, logistics order data, warehouse data, weather data, traffic data, and equipment data.

[0009] Further, the preprocessing method of the logistics multi-source heterogeneous data comprises:

[0010] The missing values in the logistics multi-source heterogeneous data are processed, the repeated data in the logistics multi-source heterogeneous data is removed, the abnormal values in the logistics multi-source heterogeneous data are processed, and preliminary standard heterogeneous data is obtained; data from different data sources in the preliminary standard heterogeneous data is merged to obtain preliminary fusion data; non-numeric data in the preliminary fusion data is converted into numeric data by encoding, and the numeric data in the preliminary fusion data is normalized / standardized to obtain preprocessed heterogeneous data.

[0011] The spatial alignment manner includes:

[0012] The original preliminary standard heterogeneous data is projected to a preliminary spatial dimension; a statistical grid is constructed on the preliminary spatial dimension, the entire data space is divided into regular grid cells, the number of data points in each grid cell is calculated, and the data distribution of each grid cell is statistically obtained; based on the number of data points in each grid cell, the data density of each grid cell is calculated, an upper limit density threshold is preset, and the grid cells higher than the upper limit density threshold are regarded as dense data regions; a lower limit density threshold is preset, and the grid cells lower than the lower limit density threshold are regarded as sparse data regions.

[0013] An initial partition granularity is defined as a basic partition of the entire data space; for the grid cells as the dense data regions, the grid cells are divided into smaller sub-cells, and the edge length of the sub-cells is wherein, ρ_min is a preset minimum density threshold, ρ is the data density of the grid cell as the dense data region; ε is a small constant; α0 is a preset initial adjustment parameter, and α1 is a linear term coefficient;

[0014] Adjacent grid cells as the sparse data regions are merged to form a plurality of partition cells.

[0015] A hierarchical structure is constructed, and the hierarchy of the hierarchical structure includes a topmost layer, an intermediate layer and a bottommost layer; the basic partition of the entire data space is regarded as the topmost layer, the sub-cells divided from the dense data regions are regarded as the intermediate layer, and the partition cells are regarded as the bottommost layer.

[0016] According to the spatial coordinates of the data points, the hierarchy to which the data points belong is determined; for the data points belonging to the intermediate layer, the data points are processed more finely; for the data points belonging to the bottommost layer, the data points are processed roughly; and then all the layers are respectively aligned according to the spatial dimension.

[0017] Further, the time alignment manner includes:

[0018] For each data source, a timestamp sequence is extracted from the preliminary standard heterogeneous data, and the extracted timestamp sequence is converted into a unified time format to obtain a unified timestamp sequence; the time interval between adjacent timestamps in the unified timestamp sequence is calculated, and the distribution characteristics of the time interval are calculated, including the mean μ, the variance σ, the median me, the lower quartile Q1, the upper quartile Q3, and the skewness sk; a uniform distribution index U is calculated using the distribution characteristics;

[0019]

[0020] where w1, w2, w3, and w4 are weight coefficients, and γ is an adjustment parameter;

[0021] A uniformity threshold is preset, and if the uniform distribution index U is greater than or equal to the uniformity threshold, then the timestamps of the preliminary standard heterogeneous data are uniformly distributed; for the uniformly distributed timestamps, linear interpolation is used for time alignment;

[0022] A number of time windows are defined, the start time points of the time windows are set, and the time windows are numbered; that is, the time windows are numbered; the unified timestamp sequence is traversed, and for each timestamp t in the unified timestamp sequence, the time window number b to which t belongs is calculated; the number of timestamps in the time window corresponding to b is counted; a time window count sequence C = [c_1, c_2, c_3,..., c_m] is obtained, where m is the total number of time windows, and c_m represents the number of timestamps in the mth time window;

[0023] A timestamp density sequence D = [d_i] is calculated based on the time window count sequence C; d_i is the density of the i th time window;

[0024] where ban is the bandwidth parameter of the kernel function, norm is the normalization coefficient; j is another index of the time window; c_j is the number of timestamps in the jth time window; w_j is the weight value of the jth time window; is an adjustment function based on the time window index j;

[0025] where len(j) is the length of the jth time window, pos(j) is the start time or center time of the jth time window, base(j) is the base function, representing the initial weight of the jth time window; α, β, and δ are adjustable parameters;

[0026] Drawing the time stamp density sequence D in time sequence into a curve, i.e. a time stamp density curve, performing peak value detection on the time stamp density curve to obtain a local maximum point p, and defining a lower threshold value x_min and an upper threshold value x_max; fixing the local maximum point p and expanding it to obtain a dense region [p-x_min, p-x_max];

[0027] For the time stamp in the dense region, spline interpolation or Gaussian kernel interpolation is used for time alignment.

[0028] Further, the mode recognition method for the preprocessed heterogeneous data comprises:

[0029] An automaton model is constructed based on the preprocessed heterogeneous data, the preprocessed heterogeneous data is mapped into a preset lattice space, i.e. the dimension and size of the lattice space are defined; the spatial coordinates in the preprocessed heterogeneous data are mapped into the lattice index of the lattice space; a set of states of the lattice in the lattice space is defined; the neighborhood radius of the lattice is defined as the Moore neighborhood, and a preset state transition rule is defined, which includes a vehicle state transition rule, an obstacle state transition rule, a weather state transition rule and a traffic state transition rule;

[0030] The initial state, time step and total iteration number of the lattice space are initialized; for each time step, all lattices are traversed, the state of the lattice at the next time step is calculated according to the current state of the lattice, the state of the neighborhood lattice and the state transition rule, and the state of all lattices is updated, and the repetition is repeated until the total iteration number is reached;

[0031] Key features are extracted from the lattice space, including clustering features, dynamic features and topological features; the clustering features are the clustering degree and clustering shape of the lattice; the dynamic features are the change trend and periodicity of the lattice state; and the topological features are the connectivity and obstacle distribution of the lattice space;

[0032] The extracted key features are input into a classification / clustering algorithm, and the lattice space is divided into different mode regions, each of which corresponds to a logistics route mode.

[0033] Further, the vehicle state transition rule is:

[0034] If the current lattice is in an idle state and there is a logistics vehicle state in the adjacent lattice, the current lattice is transferred to the corresponding logistics vehicle state with a dynamic probability PL;

[0035] PL=(1-a1-a2-a3)×P_bas+a1×f(v)+a2×g(w)+a3×h(r); wherein a1, a2 and a3 are three state weight coefficients; f(v) is a vehicle type influence function, g(w) is a weather state influence function, and h(r) is a traffic state influence function.

[0036] wherein v_max is the number of the largest vehicle type;

[0037] wherein w_max is the number of the largest weather state;

[0038] wherein r', v' and w' are positive numbers less than 1 and greater than 0, and r_max is the number of the largest traffic state;

[0039] If the current cell is a logistics vehicle state, the current logistics vehicle state is maintained with a preset maintaining probability, or is transferred to the logistics vehicle state of the adjacent cell; if the current cell is an obstacle state, the logistics vehicle state cannot be transferred to the cell.

[0040] Further, the obstacle state transition rule is:

[0041] If the current cell is an idle state, and there is an obstacle state in the adjacent cell, the current cell is transferred to the corresponding obstacle state with a preset transition probability;

[0042] If the current cell is an obstacle state, the current obstacle state is maintained, or is transferred to an idle state with an elimination probability pc;

[0043] wherein b1, b2, b3 and b4 are probability parameters, o is the number of the obstacle type; if the current cell is a logistics vehicle state or a warehouse state, the obstacle state cannot be transferred to the cell;

[0044] The weather state transition rule is:

[0045] If the current cell is an idle state, the entire research area is divided into a plurality of meteorological regions according to geographical position in advance, for each meteorological region, a spatio-temporal probability model of weather state is constructed; in the spatio-temporal probability model, given the current time and geographical position, the spatio-temporal probability model of the corresponding meteorological region is queried, the probabilities of various weather states are calculated according to historical data; the weather state with the largest probability is taken as the transition state of the current cell;

[0046] If the current cell is a weather state, the current weather state is maintained with a weather maintaining probability p_w, or is transferred to other weather states with a weather transition probability p'-w;

[0047] p'_w = (1-p_w) x P(w'|w); wherein P(w'|w) is the conditional probability that the number w of the current weather state is transferred to the number w' of the weather state;

[0048] If the current cell is a logistics vehicle state or a warehouse state, the weather state will not directly affect; if the current cell is an obstacle state, the weather state will transfer to an idle state.

[0049] Further, the preset manner of the traffic state transition rule comprises:

[0050] Model the road network as a directed weighted graph G(V, E), wherein the nodes in the node set V represent intersections or road segments, and the edges in the edge set E and their weights represent the length and capacity of the road;

[0051] Define the attributes and behavior strategies of the vehicle agent; the attributes of the vehicle agent include: current position, destination, vehicle type and travel time; the behavior strategies include path selection strategy and vehicle speed adjustment strategy; set the simulation time step dt, at each time step, all vehicle agents move in the directed weighted graph according to the behavior strategy, and the vehicle is removed from the directed weighted graph after reaching the destination, and new vehicle agents are continuously generated in the directed weighted graph according to the space-time probability model;

[0052] At each time step, count the number of vehicles on each node, and map the number of vehicles on the node to the traffic state of the node; for each node, the transition of its traffic state is affected by the traffic states of adjacent nodes, and is realized by using a rule-based transition model or a data learning-based transition model; spatially map the directed weighted graph G(V, E) and the cell space to obtain a comprehensive body; set a state transition probability matrix for the comprehensive body, and the comprehensive body with the set state transition probability matrix is the traffic state transition rule.

[0053] Further, the manner of spatially mapping the directed weighted graph G(V, E) and the cell space comprises:

[0054] Associate each cell of the cell space with a node in the directed weighted graph, and set the size of the cell to be comparable to the actual scale of the node; if there is an edge between adjacent cells, the corresponding cells also have a connection relationship; place the vehicle agent in the cell corresponding to its current position, map the attributes of the vehicle agent to different state values of the cell, and a cell can accommodate several vehicle agents; at each time step, all vehicle agents move in the cell space according to the behavior strategy, and the vehicle agent selects the next position by following the behavior strategy through the connection relationship between the cells; similarly, the vehicle agent is removed from the cell space after reaching the destination; count the number of vehicle agents in each cell, and map the number of vehicle agents to the traffic state value of the cell.

[0055] Further, the calculation manner of the optimal logistics planning scheme comprises:

[0056] Construct an optimization objective function wherein, V1, V2 and V3 are optimization weight coefficients; I and J are nodes in the directed weighted graph, (I, J) represents an edge between node I and node J, c_IJ is the unit transportation cost of edge (I, J), x_IJ is a 0-1 variable, x_IJ = 1 when the vehicle passes through edge (I, J), otherwise 0; t_IJ is the driving time of edge (I, J); h is the index of the order, SP_h is the actual delivery time of order h, QP_h is the expected delivery time of order h, N1 is the number of orders;

[0057] define optimization decision variables including vehicle scheduling scheme, route selection scheme and warehouse scheduling scheme; build constraint conditions including vehicle load, time window, driving distance and warehouse capacity, and adjust the constraint conditions according to the characteristics of M kinds of logistics route modes;

[0058] adopt permutation coding, encode the optimization decision variables into a permutation; generate an initial population according to M kinds of logistics route modes, and calculate the function value FL' of the optimization objective function of each molecule;

[0059] map the function value to the mass of the molecule wherein, ω and ω1 are positive weight parameters for adjusting the relative importance of the two terms, Fmax is the maximum function value of all molecules in the current population; PO is the penalty term of the molecule, reflecting the degree of violation of the constraint condition;

[0060] calculate the mass difference ΔM between any two molecules s1 and s2; according to the mass difference, calculate the attractive force FY_s1s2 or the repulsive force FC_s1s2 between the two molecules; for each molecule, accumulate the attractive force or repulsive force generated by all other molecules on it to obtain the total force;

[0061] for each molecule, accumulate the attractive force or repulsive force generated by all other molecules on it to obtain the total force; according to the total force, calculate the acceleration of each molecule; according to the acceleration, update the speed of the molecule; according to the speed of the molecule, update the position of the molecule; preset an upper mass threshold, mutate the molecules with mass greater than the upper mass threshold, and split them to generate new molecules; mutation is to increase / decrease the number of vehicles or change the route; the new molecule inherits the position and speed of the molecule before splitting;

[0062] preset a lower mass threshold, aggregate the molecules with mass less than the lower mass threshold; aggregation is direct replacement or fusion; the position and speed of the new molecule obtained after aggregation are calculated by weighted average according to the mass of the aggregated molecules;

[0063] The updating, mutation, splitting and aggregation are terminated until the standard deviation of the function value of the optimization objective function of the molecular group is less than a preset convergence threshold; the molecule with the largest molecular mass at this time is extracted as the optimal solution; the optimal solution is decoded into an actual logistics route planning scheme, that is, an optimal logistics planning scheme; the optimal logistics planning scheme includes vehicle scheduling, route selection and warehouse scheduling.

[0064] The technical effects and advantages of the warehouse logistics management method based on logistics big data are as follows:

[0065] The application can accurately capture the complex dynamic characteristics and potential laws of logistics data, greatly improving the accuracy of logistics prediction and decision-making. This data-driven intelligent management method not only significantly optimizes warehouse layout and inventory strategy, but also realizes dynamic optimization of transportation routes, thereby greatly reducing logistics costs and improving resource utilization efficiency. The identified logistics route patterns contain valuable experience of logistics operation, guiding subsequent planning and decision-making to avoid repeated exploration and improve efficiency. At the same time, it can realize global optimization of the entire supply chain, improve the flexibility and risk resistance of the supply chain, and the self-adaptive optimization algorithm can quickly generate the optimal solution in a complex and variable logistics environment, significantly improving logistics response speed and service quality, and improving and supplementing the original mode to form a virtuous cycle of continuous iteration and optimization. Secondly, during logistics planning, various influencing factors are considered to obtain a more comprehensive and practical optimal solution. Thirdly, the identified logistics route patterns are used as prior knowledge for optimization, greatly improving optimization efficiency and being flexible to adapt to different scales and types of logistics demand. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 FIG. 1 is a schematic diagram of a warehouse logistics management method based on logistics big data according to the present application;

[0067] Figure 2 FIG. 2 is a schematic diagram of a warehouse logistics management system based on logistics big data according to the present application. DETAILED DESCRIPTION

[0068] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0069] Embodiment 1

[0070] Please refer to Figure 1 The warehouse logistics management method based on logistics big data described in the present embodiment comprises the following steps: S1, collecting logistics multi-source heterogeneous data;

[0071] S2, preprocessing the logistics multi-source heterogeneous data to obtain preprocessed heterogeneous data;

[0072] S3, using an improved automaton algorithm to perform pattern recognition on the preprocessed heterogeneous data to obtain M kinds of logistics route patterns; M is an integer greater than 1;

[0073] S4, based on the M kinds of logistics route patterns, an optimal logistics planning scheme is obtained by optimization calculation, and the optimal logistics planning scheme is used to guide warehouse logistics management.

[0074] The logistics multi-source heterogeneous data includes logistics transportation data, logistics order data, warehouse data, weather data, traffic data and equipment data; the logistics transportation data includes vehicle GPS track data and logistics vehicle loading capacity data; the logistics order data includes customer order information, order status, shipping information and receiving information; the warehouse data includes warehouse inventory data, warehouse entry and exit records and cargo storage location;

[0075] The weather data includes weather conditions along the line, temperature and rainfall; the traffic data includes road traffic conditions, congestion conditions and construction information; the equipment data includes vehicle equipment status data and warehouse equipment status data; these heterogeneous data sources are extensive, and the data formats and structures are different, so data preprocessing and fusion are needed to provide valuable input data for subsequent feature extraction, pattern recognition and route planning.

[0076] The preprocessing method of the logistics multi-source heterogeneous data includes:

[0077] The missing values in the logistics multi-source heterogeneous data are processed, for the missing values, according to the characteristics of the data, the missing values are filled by deleting, interpolation or using statistical values (such as mean, median, etc.); the repeated data in the logistics multi-source heterogeneous data is removed, for the repeated data records, one is retained and the other repeated records are deleted; the abnormal values in the logistics multi-source heterogeneous data are processed, the abnormal values may be errors generated in the data collection, transmission or storage process, and the abnormal values need to be judged and processed according to the business rules, such as deletion, replacement, etc.; the preliminary standard heterogeneous data is obtained.

[0078] The preliminary standard heterogeneous data from different data sources is merged. Specifically, a merged time dimension and an initial space dimension are defined. The time dimension selects the smallest time granularity, such as seconds, minutes, or hours. The space dimension selects latitude and longitude coordinates or geographic coding (such as administrative division codes). Each data source in the preliminary standard heterogeneous data is time-aligned according to the time dimension, and each data source in the preliminary standard heterogeneous data is space-aligned according to the space dimension. Preliminary fusion data is obtained. Non-numerical data (such as text) in the preliminary fusion data is converted into numerical data through encoding. The numerical data in the preliminary fusion data is normalized / standardized to eliminate the influence of dimension, and preprocessed heterogeneous data is obtained.

[0079] The time alignment manner includes:

[0080] For each data source, a timestamp sequence is extracted from the preliminary standard heterogeneous data, and the extracted timestamp sequence is converted into a unified time format to obtain a unified timestamp sequence. The time interval between adjacent timestamps in the unified timestamp sequence is calculated, and the distribution characteristics of the time interval are calculated. The distribution characteristics include mean μ, variance σ, median me, lower quartile Q1, upper quartile Q3, and skewness sk. A uniform distribution index U is calculated using the distribution characteristics.

[0081]

[0082] wherein w1, w2, w3, and w4 are weight coefficients, and the numerical sum is 1, and γ is an adjustment parameter for the nonlinearity of the skewness term.

[0083] A preset uniformity threshold is set. If the uniform distribution index U is greater than or equal to the uniformity threshold, the timestamps of the preliminary standard heterogeneous data are uniformly distributed. Otherwise, they are non-uniformly distributed. For uniformly distributed timestamps, linear interpolation is used for time alignment.

[0084] A plurality of time windows are defined, such as one time window per day. The starting time point of the time window is set, such as 2023-05-01 00:00:00. The time window is numbered. That is, the time window number.

[0085] The unified timestamp sequence is traversed. For each timestamp t in the unified timestamp sequence, the time window number b to which t belongs is calculated. The number of timestamps in the time window corresponding to b is counted. A time window count sequence C = [c_1, c_2, c_3, …, c_m] is obtained, where m is the total number of time windows, and c_m represents the number of timestamps in the mth time window.

[0086] Calculate the timestamp density sequence D = [d_i] based on the time window count sequence C; d_i is the density of the i-th time window;

[0087] Where ban is the bandwidth parameter of the kernel function, controlling the smoothness of the kernel function, norm is the normalization coefficient, used to normalize the density value to the interval [0, 1]; j is another index of the time window (different from i, indexing a time window); c_j is the number of timestamps in the j-th time window; w_j is the weight value of the j-th time window; is an adjustment function based on the time window index j, used to dynamically adjust the contribution of each time window.

[0088] Where len(j) is the length of the j-th time window, pos(j) is the starting time or center time of the j-th time window, base(j) is the base function, representing the initial weight of the j-th time window, which can be a constant function or other function forms such as linear function, exponential function, etc.; α, β and δ are adjustable parameters, used to control the influence degree of time window length, position and base weight on the adjustment function.

[0089] Draw the timestamp density sequence D in time order to form a curve, which is the timestamp density curve, perform peak detection on the timestamp density curve to obtain local maximum points p, define the lower critical value x_min and the upper critical value x_max; fix the local maximum points p and expand to obtain the dense region [p-x_min, p-x_max]; use spline interpolation or Gaussian kernel interpolation for time alignment for timestamps in the dense region.

[0090] The way of spatial alignment includes:

[0091] Project the original preliminary standard heterogeneous data (such as GPS coordinates) to the preliminary spatial dimension, such as projection methods based on buffer, topological structure or probability model; construct a statistical grid in the preliminary spatial dimension, divide the entire data space (process different data sources of the preliminary standard heterogeneous data separately) into regular grid cells, calculate the number of data points in each grid cell, and statistically obtain the data distribution of each grid cell.

[0092] Based on the number of data points in each grid cell, calculate the data density of each grid cell, preset an upper limit density threshold, and regard the grid cells higher than the upper limit density threshold as dense data regions; preset a lower limit density threshold, and regard the grid cells lower than the lower limit density threshold as sparse data regions.

[0093] An initial partition granularity (e.g., road level) is defined as the basic partition of the entire data space, and for the grid cell as a dense data area, it is divided into smaller sub-cells (e.g., road segment level or finer), and the edge length of the sub-cell is where ρ_min is a preset minimum density threshold for controlling the minimum edge length of the sub-cell, ρ is the data density of the grid cell as a dense data area, ε is a small constant to avoid singular values or numerical instability of the formula when the density ρ is close to 0, and ε can be a very small normal number such as 1e-6; α0 is a preset initial adjustment parameter representing the basic value of the mapping intensity, and α1 is a linear term coefficient representing the slope of the mapping intensity with respect to the density; the two parameters α0 and α1 control the mapping intensity between the density and the sub-cell edge length in different density areas; adjacent grid cells as sparse data areas are merged to form a plurality of partition cells.

[0094] A hierarchical structure is constructed, and the hierarchy of the hierarchical structure includes a topmost layer, an intermediate layer, and a bottommost layer; the basic partition of the entire data space is taken as the topmost layer, the sub-cells divided in the dense data area are taken as the intermediate layer, and the partition cells are taken as the bottommost layer; the hierarchy to which a data point belongs is determined according to the spatial coordinates of the data point, and for the data point belonging to the intermediate layer, it is processed more finely, such as trajectory reconstruction and anomaly detection; for the data point belonging to the bottommost layer, it is processed roughly, such as simple statistics and clustering; and then all the hierarchies are respectively aligned according to the spatial dimensions (directly mapping the GPS coordinates or geographic codes to the spatial dimensions).

[0095] The mode recognition method for the preprocessed heterogeneous data includes:

[0096] An automaton model is constructed based on the preprocessed heterogeneous data, and the preprocessed heterogeneous data is mapped into a preset grid space, that is, the dimensions (two-dimensional or three-dimensional) and size (number of rows and columns, and if three-dimensional, also number of layers) of the grid space are defined; the spatial coordinates (such as latitude and longitude) in the preprocessed heterogeneous data are mapped into the grid index of the grid space (using linear interpolation, nearest neighbor interpolation, etc. to discretize the continuous spatial coordinates into the grid index).

[0097] A set of states of the grid in the grid space is defined, specifically, an idle state (0) indicating that the grid currently has no any logistics activity;

[0098] Logistics vehicle state (1, 2, 3,...): different positive integers represent different types of logistics vehicles occupying the grid, such as 1 representing a small truck, 2 representing a medium truck, 3 representing a large truck, etc.

[0099] Warehouse state (101, 102, 103,...): 100 plus a positive integer indicates that the cell is a warehouse facility, and different numbers can represent different types of warehouse facilities, such as 101 can represent a small warehouse, 102 can represent a medium-sized warehouse, 103 can represent a large warehouse, etc.

[0100] Obstacle state (-1, -2, -3,...): Different negative integers represent different types of obstacles occupying the cell, such as -1 can represent road construction, -2 can represent traffic accidents, -3 can represent natural disasters, etc.

[0101] Weather state (201, 202, 203,...): 200 plus a different positive integer indicates a different weather state, such as 201 can represent sunny, 202 can represent cloudy, 203 can represent rainy, etc.

[0102] Traffic state (301, 302, 303,...): 300 plus a different positive integer indicates a different traffic state, such as 301 can represent smooth traffic, 302 can represent light congestion, 303 can represent severe congestion, etc.

[0103] Considering multiple factors such as logistics vehicles, warehouse facilities, obstacles, weather and traffic, the automaton model can better simulate the real logistics environment; different states can be distinguished by different colors or patterns on the visualization interface.

[0104] The state set can be adjusted according to specific needs, adding or deleting some states; at the same time, different weights are assigned to each state to represent their influence on the logistics route; providing strong support for the identification of logistics route patterns.

[0105] The neighborhood radius of the cell is defined as the Moore neighborhood (8-neighborhood), and the state transition rules are preset, including vehicle state transition rules, obstacle state transition rules, weather state transition rules, and traffic state transition rules.

[0106] The vehicle state transition rule is:

[0107] If the current cell is in the idle state (0) and there is a logistics vehicle state in the adjacent cell, then with a dynamic probability PL, the current cell is transferred to the corresponding logistics vehicle state;

[0108] PL = (1-a1-a2-a3) x P_bas + a1 x f(v) + a2 x g(w) + a3 x h(r);

[0109] wherein a1, a2 and a3 are three state weight coefficients for balancing the influence of the three factors; f(v) is a vehicle type influence function, g(w) is a weather state influence function, and h(r) is a traffic state influence function, and the specific numbering is that the numbering of the vehicle type is v, v = 1 represents a small vehicle, v = 2 represents a medium vehicle, v = 3 represents a large vehicle, etc., the numbering of the weather state is w, w = 1 represents a sunny day, w = 2 represents a cloudy day, w = 3 represents a rainy day, etc., and the numbering of the traffic state is r, r = 1 represents smooth traffic, r = 2 represents light congestion, r = 3 represents severe congestion, etc.;

[0110] wherein v' is a positive number less than 1 and greater than 0, and v_max is the largest vehicle type number;

[0111] wherein w' is a positive number less than 1 and greater than 0, and w_max is the largest weather state number;

[0112] wherein r' is a positive number less than 1 and greater than 0, and r_max is the largest traffic state number.

[0113] It should be noted that the larger the vehicle type number, the higher the probability value (more likely to occupy the free grid); the worse the weather, the higher the probability value (the vehicle is more inclined to occupy the free grid); the more congested the traffic, the lower the probability value (the vehicle is difficult to occupy the free grid);

[0114] If the current grid is a logistics vehicle state, the current logistics vehicle state is maintained with a preset maintenance probability, or is transferred to the logistics vehicle state of the adjacent grid (simulating vehicle movement); if the current grid is an obstacle state, the logistics vehicle state cannot be transferred to the grid.

[0115] The obstacle state transfer rule is:

[0116] If the current grid is in an idle state (0) and there is an obstacle state in the adjacent grid, the current grid is transferred to the corresponding obstacle state with a preset transfer probability; if the current grid is in an obstacle state, the current obstacle state is maintained, or is transferred to an idle state with an elimination probability pc (simulating the duration and removal of the obstacle).

[0117] wherein b1, b2, b3 and b4 are probability parameters, b1 controls the influence degree of the obstacle type on the duration probability, b2 is a displacement parameter of the obstacle type, b3 controls the influence degree of the weather state on the duration probability, and b4 controls the influence degree of the traffic state on the duration probability; o is the number of the obstacle type (in the same way as the setting of the vehicle type); can well depict the nonlinear trend of the obstacle duration probability changing with the three influence factors.

[0118] If the current cell is in the logistics vehicle state or the storage state, the obstacle state cannot be transferred to the cell (obstacles will block vehicles and storage facilities).

[0119] The weather state transfer rule is:

[0120] If the current cell is in the idle state (0), the corresponding weather data is queried according to the current time and the geographical location of the cell to determine which weather state to transfer to; specifically, the entire study area is divided into several meteorological regions according to geographical location in advance, and for each meteorological region, a spatiotemporal probability model of weather state is constructed, such as Gaussian mixture model (GMM), Markov model, etc.; in the spatiotemporal probability model, the transfer of weather state follows certain probability distribution, and there is a transfer probability between different weather states; given the current time and geographical location, the spatiotemporal probability model of the corresponding meteorological region is queried, and the probabilities of various weather states are calculated according to historical data; the weather state with the highest probability is taken as the transfer state of the current cell.

[0121] If the current cell is in the weather state, the current weather state is maintained with a weather maintenance probability p_w (obtained from historical data), or transferred to other weather states with a weather transfer probability p'_w;

[0122] p'_w = (1-p_w) x P(w'|w); where P(w'|w) is the conditional probability of the current weather state with number w being transferred to the weather state with number w', which is given by the above model; if the current cell is in the logistics vehicle state or the storage state, the weather state will not directly affect it and will be transferred according to the normal rule; if the current cell is in the obstacle state, the weather state is transferred to the idle state.

[0123] The preset mode of the traffic state transfer rule includes:

[0124] The road network is modeled as a directed weighted graph G(V, E), where the nodes in the node set V represent intersections or road segments, and the edges and their weights in the edge set E represent the length and capacity of the road;

[0125] Define vehicle agent properties and behavior strategies; vehicle agent properties include: current location (node), destination, vehicle type and travel time; behavior strategies include path selection strategy (such as shortest path) and speed adjustment strategy (such as adjusting speed according to congestion).

[0126] Set the simulation time step dt, at each time step, all vehicle agents move in the directed weighted graph according to the behavior strategy, and the vehicle is removed from the directed weighted graph after reaching the destination, and new vehicle agents are continuously generated in the directed weighted graph according to the spatiotemporal probability model.

[0127] At each time step, the number of vehicles on each node (intersection / road segment) is counted, and the number of vehicles on the node is mapped to the traffic state of the node (e.g., free, congested, etc.). The traffic state of the node can be set to a discrete value or a continuous value (congestion level).

[0128] For each node, the transition of its traffic state is influenced by the traffic states of adjacent nodes, and a rule-based transition model or a data learning-based transition model (e.g., Markov model) is used to implement it; the transition model embodies the dynamic characteristics of traffic flow, such as spread, dispersion, periodicity, etc.

[0129] The directed weighted graph G(V, E) is spatially mapped to the grid space to obtain a complex body, each grid of the grid space is associated with a node in the directed weighted graph, and the size of the grid is set to be comparable to the actual scale of the node (scaled down) such as 100 meters x 100 meters; if there is an edge e E between adjacent grids, there is also a connection relationship between the corresponding grids; the vehicle agent is placed in the grid corresponding to its current position, and the attributes of the vehicle agent (such as vehicle type) can be mapped to different state values (discrete values) of the grid, and the same grid can accommodate several vehicle agents, indicating congestion.

[0130] At each time step, all vehicle agents move in the grid space according to the behavior strategy, and the vehicle agents select the next position by following the behavior strategy through the connection relationship between the grids, and the vehicle agents are also removed from the grid space after reaching the destination.

[0131] The number of vehicle agents in each grid is counted, and the number of vehicle agents is mapped to the traffic state value of the grid, such as 0: free; 1-10: free (specific value represents the number of vehicles): 11-20: slow; greater than 20: congested; a state transition probability matrix is set for the complex body (the traffic state transition of the grid is influenced by the traffic states of the neighboring grids; the transition probabilities in the state transition probability matrix are obtained by data learning, or some empirical rules can be set, such as the probability of transition from a congested state to a free state is lower, the probability of transition from a congested state to a higher congested state is higher, the more congested the neighboring state is, the higher the probability of transition to a congested state is; the complex body with the state transition probability matrix is the traffic state transition rule.

[0132] The traffic simulation of the agent is naturally embedded in the grid space, the movement of the agent will affect the traffic state of the grid, and the traffic state of the grid will affect the behavior strategy of the agent, and the two interact and dynamically evolve; the traffic simulation method can more microscopically and realistically depict the spatio-temporal evolution process of traffic flow, and can be organically coupled with other factors (such as logistics, weather, etc.) through the automaton model, providing strong support for the overall logistics route pattern recognition.

[0133] Initialize the initial state of the lattice space, the time step, and the total number of iterations; for each time step, iterate through all lattices, calculate the state of the lattice at the next time step according to the current state of the lattice, the state of the neighboring lattice (lattice within the Moore neighborhood), and the state transition rule, and update the state of all lattices, repeat until the total number of iterations is reached.

[0134] Extract key features from the lattice space, including clustering features, dynamic features, and topological features; clustering features include the clustering degree (the sum of the state values of all lattices within the neighborhood of the lattice) and the clustering shape (the compactness of the shape of the clustering region, i.e., the equivalent circularity index); dynamic features include the trend and periodicity of the state change of the lattice; topological features include the connectivity of the lattice space (the ratio of the number of nodes in the largest connected component to the total number of nodes) and the obstacle distribution (the proportion of lattices with obstacle states to the total number of lattices).

[0135] For each lattice, calculate the trend of its state over time, and use linear regression and other methods to fit the trend line, and the slope obtained as the trend; for each lattice, extract its time series data; use spectral analysis and other time series decomposition methods to decompose the periodic component, and calculate the goodness of fit of the periodic component to the original time series data as the periodicity; key features can reflect important information such as the spatial distribution and temporal evolution of the logistics route.

[0136] Input the extracted key features into a classification / clustering algorithm (such as support vector machine, DBSCAN clustering, K-Means clustering, etc.), divide the lattice space into different mode regions, and each mode region corresponds to a logistics route mode; simulate the dynamic evolution process of logistics data in space and time, automatically discover potential aggregation, diffusion, and other patterns in the data, and identify different logistics route modes; it has strong interpretability and can intuitively show the spatiotemporal distribution characteristics of the mode.

[0137] The calculation method of the optimal logistics planning scheme includes:

[0138] Constructing an optimization objective function where V1, V2, and V3 are optimization weight coefficients for balancing the importance of the three objectives; I and J are nodes in the directed weighted graph, (I, J) represents the edge between nodes I and J, c_IJ is the unit transportation cost of edge (I, J), x_IJ is a 0-1 variable, x_IJ = 1 when the vehicle passes through edge (I, J), otherwise 0; t_IJ is the driving time of edge (I, J); h is the index of the order, SP_h is the actual delivery time of order h, QP_h is the expected delivery time of order h, N1 is the number of orders; used to define the service level.

[0139] The definition of the optimization decision variable includes vehicle scheduling scheme (vehicle number, type), route selection scheme (selecting the road segment according to M kinds of logistics route modes) and warehouse scheduling scheme (inbound and outbound sequence).

[0140] The construction of the constraint condition includes vehicle load, time window, driving distance and warehouse capacity, and the constraint condition is adjusted according to the characteristics of the M kinds of logistics route modes; for example, analyzing the traffic condition, congestion degree and other characteristics of each kind of logistics route mode; for the mode with serious congestion, the time window constraint can be appropriately relaxed; for the mode with good road condition, the time window constraint can be appropriately reduced; the time window constraint can also be dynamically adjusted according to the weather condition, event influence and other factors.

[0141] The optimization decision variable (vehicle scheduling, route selection, warehouse scheduling) is encoded into an arrangement by using arrangement encoding; the initial molecular group (including a plurality of molecules) is generated according to the M kinds of logistics route modes, and the function value FL' of the optimization objective function of each molecule is calculated; the initial molecular group is generated in the following manner:

[0142] The feature extraction is performed on each kind of logistics route mode, including route length, number of nodes passed, average transportation time and estimated cost; an initial molecule number N3 is set, which is usually greater than M, so as to ensure sufficient diversity; a weight value is allocated to each kind of logistics route mode according to the advantages and disadvantages of the logistics route mode, and the higher the weight value, the better the mode; for each kind of logistics route mode, a plurality of initial solutions based on the logistics route mode are generated, wherein the initial solution is proportional to the weight value of the logistics route mode.

[0143] Mapping the function value into the molecular mass Wherein, ω and ω1 are positive weight parameters for adjusting the relative importance of the two items, Fmax is the maximum function value of all molecules in the current molecular group; PO is the penalty item of the molecule, reflecting the degree of violation of the constraint condition;

[0144] The mass difference ΔM between any two molecules s1 and s2 is calculated; according to the mass difference, the attractive force FY_s1s2 or the repulsive force FC_s1s2 between the two molecules is calculated; for each molecule, the attractive force or repulsive force generated by all other molecules is accumulated to obtain the total force.

[0145] Wherein, G' is the attractive force constant, R_s1s2 is the distance between two molecules s1 and s2;

[0146] Wherein, H' is the repulsive force constant, n2 is the repulsive force coefficient for controlling the strength of the repulsive force;

[0147] For each molecule, the attractive force or repulsive force generated by all other molecules is accumulated to obtain the total force.

[0148] According to the total force, the acceleration of each molecule is calculated (total force divided by the mass of the molecule); according to the acceleration, the velocity of the molecule is updated (obtained by the acceleration equation); according to the velocity of the molecule, the position of the molecule is updated (velocity time multiplied by the velocity plus the original position).

[0149] A preset upper mass threshold is set, and a molecule whose mass is greater than the upper mass threshold is mutated and split to generate a new molecule; the mutation is to increase / decrease the number of vehicles or change the route; the new molecule inherits the position and velocity of the molecule before splitting; a preset lower mass threshold is set, and a molecule whose mass is less than the lower mass threshold is aggregated to avoid falling into a local optimum; the aggregation is to directly replace or fuse; the position and velocity of the new molecule obtained after the aggregation are calculated by weighted average according to the mass of the molecules to be aggregated.

[0150] The updating, mutation, splitting, and aggregation are terminated until the standard deviation of the function value of the optimization objective function of the molecule group is less than a preset convergence threshold; the molecule with the largest mass at this time is extracted as the optimal solution; the optimal solution is decoded into an actual logistics route planning scheme, which is the optimal logistics planning scheme; the optimal logistics planning scheme includes vehicle scheduling, route selection (combined with M logistics route modes), and warehouse scheduling.

[0151] Specifically, the vehicle scheduling, route selection, and warehouse scheduling are encoded as a permutation, and the decoding steps are as follows: according to the first few bits of the encoded permutation, the scheduling information such as the number of vehicles and vehicle types is decoded; according to the middle part of the encoded permutation, the driving route of each vehicle is decoded in combination with the M logistics route modes obtained before, specifically, according to the encoding, the corresponding road sections are selected from the M logistics route modes to splice into a complete route; according to the last few bits of the encoded permutation, the warehouse scheduling information such as the order and time of warehouse entry and exit is decoded; the decoded scheme is checked to ensure that the constraint conditions are met, if not, local repair strategies such as route segmentation, customer point exchange, customer point insertion, and vehicle type upgrade are adopted; the route segmentation is to divide a long route that violates the constraints into two or more short routes; the customer point exchange is to exchange customer points between different routes of the same vehicle to make the load meet the constraints; the customer point insertion is to insert some customer points into other routes that do not violate the constraints; the vehicle type upgrade is to upgrade small vehicles to large vehicles to increase the load.

[0152] The M logistics route modes provide prior knowledge for route selection, and the modes are used as initial solutions or in the encoding, operation, boundary processing, and other links to improve performance; at the same time, the optimization process may also find new and better modes to improve and supplement the original modes.

[0153] The embodiment can accurately capture the complex dynamic characteristics and potential laws of logistics data, greatly improve the accuracy of logistics prediction and decision-making, and such a data-driven intelligent management method can not only significantly optimize warehouse layout and inventory strategy, but also realize dynamic optimization of transportation routes, thereby greatly reducing logistics costs and improving resource utilization efficiency. The identified logistics route pattern contains valuable experience of logistics operation, guiding subsequent planning and decision-making to avoid repeated exploration and improve efficiency; at the same time, it can realize global optimization of the entire supply chain, improve the flexibility and risk resistance of the supply chain, and the self-adaptive optimization algorithm can quickly generate an optimal solution in a complex and variable logistics environment, significantly improve logistics response speed and service quality, and improve and supplement the original mode to form a virtuous cycle of continuous iteration and optimization; secondly, in the logistics planning, various influencing factors are considered to obtain a more comprehensive and actual optimal solution; thirdly, the identified logistics route pattern is used as priori knowledge for optimization, which greatly improves the optimization efficiency and can adapt to different scales and types of logistics demand.

[0154] Embodiment 2

[0155] Please refer to Figure 2 The embodiment does not describe some parts in detail, see the description of embodiment 1, and provides a warehouse logistics management system based on logistics big data, comprising: a data acquisition module for acquiring logistics multi-source heterogeneous data;

[0156] A data processing module is configured to preprocess the logistics multi-source heterogeneous data to obtain preprocessed heterogeneous data.

[0157] A pattern recognition module is configured to recognize M logistics route patterns from the preprocessed heterogeneous data by using an improved automaton algorithm.

[0158] A dynamic programming module is configured to obtain an optimal logistics planning scheme by optimizing calculation based on the M logistics route patterns, and to guide warehouse logistics management by using the optimal logistics planning scheme. The modules are connected by wired and / or wireless means to realize data transmission between the modules.

[0159] Embodiment 3

[0160] The embodiment discloses an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to realize the operation mode of the warehouse logistics management method based on logistics big data provided above.

[0161] Since the electronic device introduced in the embodiment is the electronic device used in the implementation of the warehouse logistics management method based on logistics big data in the embodiment of the application, based on the warehouse logistics management method based on logistics big data in the embodiment of the application, those skilled in the art can understand the specific implementation of the electronic device of the embodiment and its various forms, so the implementation of the method in the embodiment of the application is not introduced in detail. As long as the electronic device used in the implementation of the warehouse logistics management method based on logistics big data in the embodiment of the application belongs to the scope of protection of the application.

[0162] The above formulas are dimensionless values, and the formulas are obtained by collecting a large amount of data to simulate the most recent real situation. The preset parameters and threshold values in the formula are set by those skilled in the art according to the actual situation.

[0163] The above is only the preferred embodiment of the application, and the protection scope of the application is not limited to the above-mentioned embodiments. Any technical solution that belongs to the idea of the application is within the protection scope of the application. It should be noted that for ordinary technical users in the technical field, some improvements and decorations without departing from the principles of the application are also considered as the protection scope of the application.

Claims

1. A warehousing logistics management method based on logistics big data, characterized in that: include: S1. Collecting multi-source heterogeneous logistics data; the multi-source heterogeneous logistics data includes logistics transportation data, logistics order data, warehousing data, weather data, traffic data and equipment data S2. Preprocessing the logistics multi-source heterogeneous data to obtain preprocessed heterogeneous data; S3, using the improved automaton algorithm to perform pattern recognition on the pre-processed heterogeneous data and obtain M types of logistics route patterns; The method of performing pattern recognition on pre-processed heterogeneous data includes: An automaton model is constructed based on preprocessed heterogeneous data, and the preprocessed heterogeneous data is mapped into a preset grid space, that is, the dimension and size of the grid space are defined; the spatial coordinates in the preprocessed heterogeneous data are mapped to the grid index of the grid space; the set of grid states in the grid space is defined; the neighborhood radius of the grid is defined as the Moore neighborhood, and state transition rules are preset. The state transition rules include vehicle state transition rules, obstacle state transition rules, weather state transition rules, and traffic state transition rules; Initialize the initial state, time step, and total number of iterations of the grid space; for each time step, traverse all grids, calculate the state of the grid in the next time step based on the state of the current grid, the state of the neighboring grids, and the state transition rule, and update the state of all grids, repeating until the total number of iterations is reached; Extract key features from the grid space, including clustering features, dynamic features, and topological features; clustering features are the clustering degree and cluster shape of the grid; dynamic features are the changing trend and periodicity of the grid state; topological features are the connectivity and obstacle distribution of the grid space; The extracted key features are input into the classification / clustering algorithm to divide the grid space into different pattern areas, each of which corresponds to a logistics route pattern; S4. Based on M types of logistics route patterns, optimize and calculate to obtain the optimal logistics planning scheme, and use the optimal logistics planning scheme to guide warehouse logistics management.

2. A warehousing logistics management method based on logistics big data according to claim 1, characterized in that: The method of preprocessing the logistics multi-source heterogeneous data includes: Process missing values ​​in logistics multi-source heterogeneous data, remove duplicate data in logistics multi-source heterogeneous data, process outliers in logistics multi-source heterogeneous data, and obtain preliminary standard heterogeneous data; merge data from different data sources in the preliminary standard heterogeneous data to obtain preliminary fused data; specifically, align each data source in the preliminary standard heterogeneous data according to the time dimension, and align each data source in the preliminary standard heterogeneous data according to the spatial dimension to obtain preliminary fused data; encode non-numeric data in the preliminary fused data into numeric data, and normalize / standardize the numeric data in the preliminary fused data to obtain preprocessed heterogeneous data; The spatial alignment method includes: The original preliminary standard heterogeneous data is projected onto the preliminary spatial dimension; a statistical grid is constructed on the preliminary spatial dimension, the entire data space is divided into regular grid cells, the number of data points in each grid cell is calculated, and the data distribution of each grid cell is obtained statistically; based on the number of data points in each grid cell, the data density of each grid cell is calculated, an upper density threshold is preset, and grid cells above the upper density threshold are regarded as dense data areas; a lower density threshold is preset, and grid cells below the lower density threshold are regarded as sparse data areas; Define an initial partition granularity as the basic partition of the entire data space. For the grid cells that are dense data areas, divide them into smaller sub-cells with a side length of ;in, is a preset minimum density threshold, is the data density of the grid cells that are dense data areas; is a constant; is the preset initial adjustment parameter, is the linear term coefficient; Merge adjacent grid cells that are sparse data areas to form several partition units; Construct a hierarchical structure with the top, middle, and bottom layers. The basic partition of the entire data space is the top layer, the sub-units of the dense data area are the middle layer, and the partition units are the bottom layer. According to the spatial coordinates of the data points, the layer to which they belong is determined. For the data points belonging to the middle layer, they are processed more finely; for the data points belonging to the bottom layer, they are processed roughly; and then all layers are aligned according to the spatial dimensions.

3. A warehousing logistics management method based on logistics big data according to claim 2, characterized in that: The time alignment method includes: For each data source, extract the timestamp sequence from the preliminary standard heterogeneous data and convert the extracted timestamp sequence into a unified time format to obtain a unified timestamp sequence; calculate the time intervals between adjacent timestamps in the unified timestamp sequence and calculate the distribution characteristics of the time intervals, including the mean ,variance , median , lower quartile , upper quartile and skewness ;Use the distribution characteristics to calculate the uniform distribution index ; ;in, 、 、 and is the weight coefficient, To adjust the parameters; Preset uniformity threshold, if uniform distribution index When it is greater than or equal to the uniformity threshold, the timestamps of the preliminary standard heterogeneous data are uniformly distributed; for uniformly distributed timestamps, linear interpolation is used for time alignment; Define several time windows, set the starting time points of the time windows, and number the time windows; this is the time window number; traverse the unified timestamp sequence, and for each timestamp t in the unified timestamp sequence, calculate the time window number b to which t belongs; count the number of timestamps in the time window corresponding to b; obtain the time window count sequence C=[c_1, c_2, c_3, ..., c_m], where m is the total number of time windows and c_m represents the number of timestamps in the mth time window; The timestamp density sequence D = [d_i] is calculated based on the time window count sequence C; For the The density of the time windows; ;in, is the bandwidth parameter of the kernel function, is the normalization coefficient; is another index for the time window; For the The number of timestamps in a time window; For the The weight value of a time window; For a time window index The adjustment function of ;in, For the The length of the time window, For the The start time or center time of a time window, is the basic function, indicating the The initial weight of the time window; 、 and It is an adjustable parameter; The timestamp density sequence D is plotted into a curve in chronological order, which is the timestamp density curve. The timestamp density curve is peak detected to obtain the local maximum point p, and the lower critical value x_min and the upper critical value x_max are defined. The local maximum point p is fixed and expanded to obtain the dense area [p-x_min, p-x_max]. For timestamps in dense areas, spline interpolation or Gaussian kernel interpolation is used for time alignment.

4. A warehousing logistics management method based on logistics big data according to claim 3, characterized in that: The vehicle state transition rule is: If the current grid is idle and there is a logistics vehicle in the adjacent grid, the dynamic probability Transfer the current grid to the corresponding logistics vehicle state; ;in, 、 and are the three state weight coefficients; is the vehicle type influence function, is the weather status impact function, is the traffic state impact function; ;in, The number for the largest vehicle type; ;in, is the number of the maximum weather state; in, 、 and are all positive numbers less than 1 and greater than 0. is the number of the largest traffic state; If the current grid is in the logistics vehicle state, the current logistics vehicle state will be maintained with a preset retention probability, or transferred to the logistics vehicle state of an adjacent grid; if the current grid is in the obstacle state, the logistics vehicle state cannot be transferred to the grid.

5. A warehousing and logistics management method based on logistics big data according to claim 4, characterized in that: The obstacle state transfer rule is: If the current grid is in an idle state and there is an obstacle state in the adjacent grid, the current grid is transferred to the corresponding obstacle state with the preset transition probability; If the current grid is in an obstacle state, keep the current obstacle state, or eliminate the probability Transfer to idle state; ;in, 、 、 and is the probability parameter, The number of the obstacle type. If the current grid is in the logistics vehicle state or the warehouse state, the obstacle state cannot be transferred to this grid. The weather state transition rules are: If the current grid is idle, pre-divide the entire study area into several meteorological zones based on their geographical locations. For each meteorological zone, construct a spatiotemporal probability model of weather conditions. Given the current time and geographical location, query the spatiotemporal probability model of the corresponding meteorological zone and calculate the probabilities of various weather conditions based on historical data. The weather condition with the highest probability is used as the transition state for the current grid. If the current grid is in weather state, the weather maintains the probability Maintain the current weather status, or change the weather status based on the probability of weather transition Transfer to other weather conditions; ;in, The number of the current weather status Number of weather status transferred to The conditional probability of If the current grid is in a logistics vehicle state or a warehouse state, the weather state will not have a direct impact; if the current grid is in an obstacle state, the weather state will be transferred to an idle state.

6. A warehousing logistics management method based on logistics big data according to claim 5, characterized in that: The preset methods of traffic state transition rules include: The road network is modeled as a directed weighted graph G(V, E), where the nodes in the node set V represent intersections or road sections, and the edges in the edge set E and their weights represent the length and capacity of the road; Define vehicle agent attributes and behavior strategies. Vehicle agent attributes include current location, destination, vehicle model, and travel time. Behavior strategies include path selection and speed control strategies. Set the simulation time step dt. At each time step, all vehicle agents move in the directed weighted graph according to the behavior strategy. After reaching the destination, the vehicle is removed from the directed weighted graph. New vehicle agents are continuously generated in the directed weighted graph based on the spatiotemporal probability model. At each time step, the number of vehicles at each node is counted and mapped to the node's traffic state. For each node, the transition of its traffic state is affected by the traffic state of adjacent nodes and is implemented using a rule-based transition model or a data learning-based transition model. The directed weighted graph G(V, E) is spatially mapped to the grid space to obtain a complex. A state transition probability matrix is ​​set for the complex, and the complex with the state transition probability matrix is ​​the traffic state transition rule.

7. A warehousing logistics management method based on logistics big data according to claim 6, characterized in that: The method of spatially mapping the directed weighted graph G(V, E) to the grid space includes: Each grid in the grid space is associated with a node in a directed weighted graph, and the size of the grid is set to be equivalent to the actual scale of the node; if there is an edge between adjacent grids, the corresponding grids are also connected; the vehicle agent is placed in the grid corresponding to its current position, and the attributes of the vehicle agent are mapped to different state values ​​of the grid. The same grid can accommodate several vehicle agents; at each time step, all vehicle agents move in the grid space according to the behavioral strategy. The vehicle agent selects the next position according to the behavioral strategy through the connectivity between grids. Similarly, the vehicle agent is removed from the grid space after reaching the destination; the number of vehicle agents in each grid is counted, and the number of vehicle agents is mapped to the traffic state value of the grid.

8. A warehousing logistics management method based on logistics big data according to claim 7, characterized in that: The calculation method of the optimal logistics planning solution includes: Constructing the optimization objective function ;in, 、 and To optimize the weight coefficient; and is a node in a directed weighted graph, Representation node and nodes The edges between For the edge The unit transportation cost, It is a 0-1 variable. When the vehicle passes the edge hour , otherwise 0; For the edge driving time; is the index of the order, For orders The actual delivery time, For orders Expected delivery time, is the number of orders; The optimization decision variables are defined, including vehicle scheduling, route selection, and warehouse scheduling. Constraints are constructed, including vehicle load, time window, driving distance, and warehouse capacity, and the constraints are adjusted according to the characteristics of M logistics route patterns. Using permutation coding, the optimization decision variables are encoded into a permutation; based on M types of logistics route patterns, an initial molecule group is generated, and the function value of the optimization objective function of each molecule is calculated. ; Mapping function values ​​to molecular masses ;in, and is a positive weight parameter used to adjust the relative importance of the two items. is the maximum function value of all molecules in the current molecular group; is the penalty term of the numerator, reflecting the degree of violation of the constraint conditions; Calculate the mass difference between any two molecules s1 and s2 ; Calculate the attraction between two molecules based on the mass difference or repulsion For each molecule, add up the attractive or repulsive forces exerted on it by all other molecules to get the total force; Based on the total force, the acceleration of each molecule is calculated; based on the acceleration, the molecular velocity is updated; based on the molecular velocity, the molecular position is updated; an upper mass threshold is preset, and molecules with a molecular mass greater than the upper mass threshold are mutated and split to produce new molecules; mutations are performed to increase / decrease the number of vehicles or change the route; the new molecules inherit the position and velocity of the molecules before the split; A lower mass threshold is preset, and molecules with a mass less than the lower mass threshold are polymerized; the polymerization is performed as direct replacement or fusion; the position and velocity of the new molecules obtained after polymerization are calculated based on the weighted average of the mass of the polymerized molecules; Until the standard deviation of the function value of the optimization objective function of the molecular group is less than the preset convergence threshold, the update, mutation, split and aggregation are terminated; the molecule with the largest molecular mass at this time is extracted as the optimal solution; the optimal solution is decoded into the actual logistics route planning plan, which is the optimal logistics planning plan; the optimal logistics planning plan includes vehicle scheduling, route selection and warehouse scheduling.

Images