Multi-dimensional dynamic fusion intelligent logistics scheduling method and system
By quantifying and converting logistics tasks and transportation resources and analyzing historical data, and by generating scheduling strategies in conjunction with abnormal situations, the problem of logistics scheduling being unable to adapt to complex scenarios has been solved, and highly accurate and flexible logistics scheduling has been achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HANGZHOU GOLDEN SOFTWARE SYST INC
- Filing Date
- 2026-04-02
- Publication Date
- 2026-05-12
AI Technical Summary
Existing logistics scheduling methods cannot adapt to the full-dimensional needs of complex logistics scenarios, mainly because they only focus on time or cost indicators and cannot meet the diverse needs of different logistics orders.
By collecting logistics task parameters and transportation resource parameters, quantitative conversion is performed to generate task quantitative values and transportation resource quantitative values. Combined with historical scheduling data, the optimal matching scheduling strategy is determined, and scheduling adjustment strategies are generated based on abnormal situations, thereby realizing the fusion and dynamic adaptation of multi-dimensional parameters.
It significantly improves the accuracy, flexibility and reliability of logistics scheduling, and can respond to real-time tasks and sudden anomalies, adapting to the full-dimensional needs of complex logistics scenarios.
Smart Images
Figure CN122022385A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of logistics scheduling technology, and in particular to a multi-dimensional dynamic fusion intelligent logistics scheduling method and system. Background Technology
[0002] Logistics scheduling refers to the work of coordinating and allocating various resources across the entire logistics chain and coordinating the operational processes of each link based on actual operational needs, so as to ensure that goods are transferred accurately, timely, and at low cost as required.
[0003] Currently, when conducting logistics scheduling, we generally first receive logistics orders from procurement, production, sales, and other stages, sort out basic information such as the transportation route, time requirements, material attributes, and budget range of the goods in the order, and then, based on time or cost indicators, select vehicles and carriers that meet the basic vehicle type and load requirements from the transportation capacity resources, formulate fixed transportation routes and loading plans, and carry out scheduling.
[0004] Because different logistics orders have different compliance conditions and different service quality requirements, and current logistics scheduling generally only focuses on time or cost indicators, it cannot adapt to the full-dimensional needs of complex logistics scenarios. Summary of the Invention
[0005] To facilitate logistics scheduling and adapt to the full range of needs in complex logistics scenarios, this invention provides a multi-dimensional dynamic fusion intelligent logistics scheduling method and system.
[0006] Firstly, the present invention provides a multi-dimensional dynamic fusion intelligent logistics scheduling method, which adopts the following technical solution:
[0007] A multi-dimensional dynamic fusion intelligent logistics scheduling method includes:
[0008] Collect logistics task parameters, transportation capacity resource parameters, and historical scheduling data;
[0009] The task parameters are quantified and converted to generate a task quantification value, and the transportation capacity resource parameters are quantified and converted to generate a transportation capacity quantification value.
[0010] The optimal matching scheduling strategy is determined by combining historical scheduling data, task quantification values, and transportation capacity quantification values.
[0011] Dynamically optimize the scheduling strategy based on the optimal matching strategy to generate an optimized scheduling strategy;
[0012] Logistics scheduling is carried out based on optimized scheduling strategies, and parameters for abnormal situations are collected.
[0013] The scheduling adjustment strategy is determined by combining abnormal situation parameters with the optimized scheduling strategy, and then the scheduling adjustment strategy is executed.
[0014] By adopting the above technical solution, logistics task parameters, transportation capacity resource parameters, and historical scheduling data are collected, and task quantitative values and transportation capacity quantitative values are generated through quantification and transformation. The optimal matching scheduling strategy is determined and dynamically optimized by combining the three types of data. At the same time, scheduling adjustment strategies are generated based on abnormal situation parameters. This achieves comprehensive integration of multi-dimensional parameters and dynamic adaptation of the entire scheduling process, effectively avoiding the limitations of traditional scheduling that relies on a single parameter. It ensures that logistics scheduling not only conforms to historical scheduling experience but also responds to real-time task, transportation capacity changes, and sudden anomalies. This facilitates logistics scheduling to adapt to the full-dimensional needs of complex logistics scenarios and significantly improves the accuracy, flexibility, and reliability of scheduling.
[0015] Optional, the specific steps of quantization conversion include:
[0016] Retrieve parameter types and data based on logistics task parameters or transportation capacity resource parameters;
[0017] The conversion type is determined based on the type of parameters. Conversion types include hierarchical encoding type, normalization type, and Boolean encoding type.
[0018] Transform category data based on transformation type.
[0019] By adopting the above technical solution, the parameter types and data of logistics task parameters or transportation capacity resource parameters are first retrieved, and then the parameter types are converted according to the appropriate hierarchical coding type, normalization type or Boolean coding type. This achieves precise matching between quantitative conversion and parameter characteristics, providing standardized and high-quality data support for determining the optimal matching scheduling strategy.
[0020] Optionally, transforming categorical data based on the transformation type includes:
[0021] Retrieve the maximum, minimum, and real-time values of data based on the category data;
[0022] When the conversion type is normalization, the difference between the maximum and minimum data values is calculated and used as the data baseline deviation value.
[0023] Calculate the difference between the real-time data value and the minimum data value and use it as the real-time deviation value of the data;
[0024] The conversion is performed based on the ratio between the real-time deviation value and the baseline deviation value of the data.
[0025] When the conversion type is Boolean encoding, the category baseline data is retrieved based on the parameter category, and the conversion is performed based on the matching result between the category data and the category baseline data.
[0026] By adopting the above technical solutions, the conversion of normalized data is achieved by calculating the ratio of the data baseline deviation value to the data real-time deviation value, ensuring that different types of data are mapped to a unified range and improving data comparability. For Boolean encoded data, the conversion is achieved by matching the type data with the type baseline data, clarifying the binary attribute of the parameters. Both conversion methods ensure the accuracy and standardization of the quantized values, providing a precise data foundation for the calculation of subsequent scheduling strategies.
[0027] Optionally, transforming categorical data based on the transformation type also includes:
[0028] When the conversion type is hierarchical encoding, the relevant type and related data are retrieved based on the parameter type.
[0029] The correlation coefficient is determined based on the relevant categories, and a single reference value is determined in combination with the relevant data;
[0030] A comprehensive reference value is obtained by weighting the correlation coefficient with the relevant single reference value.
[0031] The reference level range is retrieved based on the parameter type and matched with relevant comprehensive reference values for conversion.
[0032] By adopting the above technical solution, for the hierarchical coding type, by retrieving relevant categories and data, determining relevant single reference values, calculating relevant comprehensive reference values, and then matching the reference level range to complete the conversion, the comprehensive quantification of parameters affected by multiple factors is realized, so that the task quantification value or operation force quantification value is adapted to the scheduling scenario that requires hierarchical evaluation, and the comprehensiveness and pertinence of quantification conversion are improved.
[0033] Optionally, methods for determining the optimal matching scheduling strategy include:
[0034] Determine the basic weights and trigger condition set based on historical scheduling data;
[0035] The fluctuation adjustment value is determined by matching and comparing the task quantification value, the transportation quantification value, and the set of triggering conditions.
[0036] The base weights are adjusted based on the fluctuation adjustment value to obtain the adjusted weights;
[0037] The adjusted weights, task quantification values, and transportation quantification values are input into a preset integer programming model to obtain a planning and scheduling strategy, which is then used as the optimal matching scheduling strategy.
[0038] By adopting the above technical solution, the basic weights and trigger condition set are determined through historical scheduling data. The fluctuation adjustment value is calculated by combining the task quantification value and the transportation force quantification value. The adjustment weight is then calculated. By inputting the adjustment weight, task quantification value, and transportation force quantification value into the integer programming model, the planning scheduling strategy is obtained. This realizes the dynamic adaptation of weights and the scientific planning of scheduling strategies, ensuring that the optimal matching scheduling strategy can both inherit the effective experience of historical scheduling and respond to real-time parameter fluctuations, thereby improving the rationality and adaptability of the scheduling strategy.
[0039] Optional methods for determining the basic weights and trigger condition set include:
[0040] Retrieve scheduling sample types and individual sample data based on historical scheduling data;
[0041] Determine the maximum deviation value, median value, and reference value of the distribution based on single sample data;
[0042] The trigger selection value is determined based on the distribution level reference value;
[0043] The numerical value of each category is determined based on the category of the scheduling sample;
[0044] The weight of a single category is determined by combining the numerical value of the number of categories with the reference value of the degree of distribution.
[0045] Calculate the product of the selected trigger value and the maximum distribution deviation value, and use it as the single trigger deviation value of the distribution.
[0046] The basic weight is the single weight corresponding to each type of scheduling sample, and the trigger condition set is the distribution median value and the single trigger deviation value corresponding to each type of scheduling sample.
[0047] By adopting the above technical solution, the maximum deviation value, median value, and reference value of distribution degree are determined through historical scheduling data. Then, the single-type weight and trigger condition set are determined, so that the basic weight can reflect the influence priority of different types of scheduling samples, and the trigger condition set fits the data distribution characteristics. This provides an objective and reliable basis for subsequent fluctuation adjustment value calculation and optimal matching scheduling strategy determination, and improves the scientific nature of scheduling decisions.
[0048] Optionally, methods for determining the fluctuation adjustment value include:
[0049] Based on the task quantification value or the transportation force quantification value, select the corresponding distribution median value from the set of triggering conditions and use it as the selection benchmark value;
[0050] Define a task quantification value or a transportation quantification value as the current selection value, and retrieve the current selection type based on the current selection value;
[0051] Calculate the difference between the current selected value and the selected reference value and use it as the selection deviation value;
[0052] Calculate the ratio between the selected deviation value and the selected benchmark value, and use it as the selection ratio value;
[0053] Determine the sensitivity coefficient of the species based on the currently selected species;
[0054] Calculate the product between the selected proportion value and the type sensitivity coefficient, and use it as the fluctuation adjustment value.
[0055] By adopting the above technical solution, the selection deviation value and selection ratio value are calculated by selecting a benchmark value and the current selected value, and the fluctuation adjustment value is obtained by combining the type sensitivity coefficient. This makes the fluctuation adjustment value affected by the actual fluctuation degree of the parameter and the importance of the parameter, ensuring that the optimal matching scheduling strategy can adapt to parameter fluctuations in a timely manner and improving the dynamic response capability of the scheduling strategy.
[0056] Optionally, methods for generating optimized scheduling strategies include:
[0057] Selecting and optimizing algorithm types for data acquisition;
[0058] Retrieve numerical values from a single sample of data;
[0059] The algorithm configuration parameters are determined by combining the selection of algorithm types, the number of samples, and the number of types.
[0060] Based on the algorithm configuration parameters, the algorithm corresponding to the selected algorithm type is configured and used as the optimization configuration algorithm;
[0061] The optimal matching scheduling strategy is input into the optimization configuration algorithm to obtain the optimized scheduling strategy.
[0062] By adopting the above technical solution, the types of algorithms are selected through data collection and optimization. The algorithm configuration parameters are determined by combining the selected algorithm types, the number of samples, and the number of types. After configuring the algorithm, the optimal matching scheduling strategy is optimized, which further improves the optimality and pertinence of the scheduling strategy.
[0063] Optionally, methods for determining the scheduling adjustment strategy include:
[0064] Retrieve the anomaly type and its impact range based on the abnormal situation parameters;
[0065] The response level is determined by combining the anomaly type and the scope of its impact;
[0066] The adjustment type is determined based on the response level;
[0067] The intervention and adjustment strategy is determined by combining the adjustment type and the abnormality type;
[0068] The optimal scheduling strategy is adjusted based on the intervention and adjustment strategy, and the scheduling adjustment strategy is obtained.
[0069] By adopting the above technical solution, the abnormal situation parameters are used to retrieve the abnormality type and scope of impact, determine the adjustment type, generate intervention and adjustment strategies, and adjust and optimize the scheduling strategy. This achieves hierarchical and precise response to abnormalities, thereby taking appropriate adjustment measures according to the severity of the abnormality, quickly resolving the impact of the abnormality, and ensuring the continuity and stability of logistics scheduling.
[0070] Secondly, this invention provides a multi-dimensional dynamic fusion intelligent logistics scheduling system, which adopts the following technical solution:
[0071] A multi-dimensional dynamic fusion intelligent logistics scheduling system includes:
[0072] The data acquisition module is used to collect logistics task parameters, transportation capacity resource parameters, historical scheduling data, abnormal situation parameters, and to select optimization algorithms.
[0073] The memory stores a program for implementing a multi-dimensional dynamic fusion intelligent logistics scheduling method as described in any one of the first aspects;
[0074] The processor loads and executes programs stored in memory.
[0075] In summary, the present invention has at least one of the following beneficial technical effects:
[0076] 1. The system collects logistics task parameters, transportation capacity parameters, and historical scheduling data. After quantification and transformation, it generates task quantification values and transportation capacity quantification values. Combining these three types of data, it determines the optimal matching scheduling strategy and dynamically optimizes it. At the same time, it generates scheduling adjustment strategies based on abnormal situation parameters. This achieves comprehensive integration of multi-dimensional parameters and dynamic adaptation of the entire scheduling process, effectively avoiding the limitations of traditional scheduling that relies on a single parameter. It ensures that logistics scheduling not only conforms to historical scheduling experience but also responds to real-time task, transportation capacity changes, and sudden anomalies. This facilitates logistics scheduling to adapt to the full-dimensional needs of complex logistics scenarios and significantly improves the accuracy, flexibility, and reliability of scheduling.
[0077] 2. First, retrieve the parameter types and data of logistics task parameters or transportation resource parameters, and then convert them according to the parameter type by matching hierarchical coding type, normalization type or Boolean coding type. This achieves precise matching between quantitative conversion and parameter characteristics, providing standardized and high-quality data support for determining the optimal matching scheduling strategy in the future.
[0078] 3. By retrieving the anomaly type and its impact range through abnormal situation parameters, the adjustment type is determined, and then intervention and adjustment strategies are generated and the scheduling strategy is adjusted and optimized. This achieves hierarchical and precise response to anomalies, thereby taking appropriate adjustment measures according to the severity of the anomaly, quickly resolving the impact of the anomaly, and ensuring the continuity and stability of logistics scheduling. Attached Figure Description
[0079] Figure 1 This is a flowchart of a multi-dimensional, dynamic, integrated intelligent logistics scheduling method. Detailed Implementation
[0080] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.
[0081] A multi-dimensional dynamic fusion intelligent logistics scheduling method is proposed. This method collects logistics task parameters, transportation capacity parameters, historical scheduling data, anomaly parameters, and optimization algorithm types. Quantization transformation is performed according to parameter types using hierarchical coding, normalization, and Boolean coding. Based on historical data, basic weights and triggering conditions are determined. Fluctuation adjustment values and adjustment weights are calculated using the quantified values. An optimal matching scheduling strategy is obtained through an integer programming model. Subsequently, the configured optimization algorithm types are used to optimize and generate an optimized scheduling strategy for logistics scheduling. During the scheduling process, anomaly parameters are collected, and intervention and adjustment strategies are determined and executed hierarchically according to the anomaly type and its impact range. This ensures that logistics scheduling not only conforms to historical scheduling experience but also responds to real-time tasks, transportation capacity changes, and sudden anomalies, facilitating the adaptation of logistics scheduling to the full-dimensional needs of complex logistics scenarios.
[0082] Reference Figure 1 This invention discloses a multi-dimensional dynamic fusion intelligent logistics scheduling method, which includes:
[0083] S1: Collect logistics task parameters, transportation capacity resource parameters, and historical scheduling data.
[0084] Logistics task parameters refer to various attribute information directly related to logistics and transportation tasks. These parameters can be obtained through order system integration, manual entry, and other methods. They include route plans (origin and destination, transit nodes, path constraints), task type (procurement / production / sales / reverse), time requirements (delivery window, urgency, SLA requirements), vehicle requirements (vehicle type, load capacity, special equipment), material requirements (category, specifications, special handling, dangerous goods / fragile goods labeling), task urgency, carrier rating requirements, task budget range, and task KPI assessment parameters (on-time delivery rate, damage rate, etc.).
[0085] Transportation capacity resource parameters refer to the attribute information of various resources such as carriers and vehicles undertaking logistics tasks. These parameters are obtained through transportation capacity management systems, GPS devices, and carrier reporting. These parameters include carrier type (owned / third-party / crowdsourced / dedicated line), carrier level (qualifications, service capabilities, historical performance), vehicle type (van / flatbed / refrigerated truck, etc.), vehicle status (idle / en route / under maintenance / standby), transport type (general cargo / refrigerated / dangerous goods), vehicle location (latitude and longitude, regional distribution), response time (time from order acceptance to departure), transportation cost (base price, mileage fee, surcharges), historical performance (on-time rate, complaint rate, cargo damage rate), and equipment configuration (GPS, temperature control, security equipment).
[0086] Historical scheduling data refers to various records generated from past completed logistics scheduling operations. It includes information such as tasks, transportation capacity, execution process, and results. Historical scheduling data is extracted from the scheduling system's database.
[0087] S2: Quantify and convert logistics task parameters to generate task quantification values, and quantify and convert transportation capacity resource parameters to generate transportation capacity quantification values.
[0088] Quantification transformation refers to the process of converting unstructured or scattered parameter information into standardized, computable numerical forms. Task quantification value refers to the standardized value obtained after quantification transformation of logistics task parameters. Transportation capacity quantification value refers to the standardized value obtained after quantification transformation of transportation capacity resource parameters.
[0089] By quantifying and converting the logistics task parameters and transportation capacity resource parameters respectively, we can obtain the quantified values of the task and transportation capacity, which will be convenient for subsequent use.
[0090] To further ensure the rationality of the quantification conversion, a more detailed separate analysis and calculation of the quantification conversion is required, which will be explained in detail through the steps shown below.
[0091] The specific steps of quantization conversion include the following:
[0092] S21: Retrieve parameter types and data based on logistics task parameters or transportation capacity resource parameters.
[0093] Here, parameter type refers to the specific categories of logistics task parameters or transportation capacity resource parameters based on their attribute characteristics. Category data refers to the specific information or values under the corresponding parameter type. Both logistics task parameters and transportation capacity resource parameters include parameter types and category data.
[0094] The parameters and their data can be retrieved by using logistics task parameters or transportation capacity resource parameters, which will facilitate subsequent use.
[0095] S22: Determine the conversion type based on the parameter type.
[0096] Here, conversion type refers to the specific conversion rule category preset to achieve parameter standardization and quantization. Conversion types include hierarchical encoding type, normalization type, and Boolean encoding type.
[0097] By inputting the parameter types into a preset type conversion database, the conversion type can be matched and obtained for convenient subsequent use.
[0098] The category conversion database pre-stores a table of different parameter types and their corresponding conversion types, which is retrieved after the operator pre-inputs the information.
[0099] For example, when the parameter type is task urgency or carrier level, the corresponding conversion type is hierarchical coding; when the parameter type is time requirement, task budget range, or response time, the corresponding conversion type is normalization; and when the parameter type is vehicle status or transportation type, the corresponding conversion type is Boolean coding.
[0100] S23: Transform category data based on transformation type.
[0101] In this process, the data corresponding to each parameter type is converted by conversion type to obtain standardized values, which are convenient for subsequent use.
[0102] To further ensure the rationality of transforming category data based on transformation type, it is necessary to perform further separate analysis and calculation on the transformation of category data based on transformation type, which will be explained in detail through the following steps.
[0103] Transforming category data based on transformation type includes the following steps:
[0104] S231: Retrieve the maximum value, minimum value, and real-time value of data based on the category data.
[0105] Among these, the maximum value refers to the largest value among the category data under the same parameter category. The minimum value refers to the smallest value among the category data under the same parameter category. The real-time value refers to the specific actual data corresponding to the parameter category to be quantized and transformed. Category data includes the maximum value, minimum value, and real-time value.
[0106] The maximum, minimum, and real-time values of the data can be retrieved by category data for convenient subsequent use.
[0107] S232: When the conversion type is normalization, calculate the difference between the maximum and minimum data values and use it as the data baseline deviation value.
[0108] Among them, the data benchmark deviation value refers to the range of numerical fluctuation of the data under the same parameter category.
[0109] When the conversion type is normalization, the difference between the maximum and minimum data values is calculated, and the result is used as the data baseline deviation value for convenient subsequent use.
[0110] S233: Calculate the difference between the real-time value of the data and the minimum value of the data, and use it as the real-time deviation value of the data.
[0111] The real-time deviation value refers to the difference between the real-time value of the data and the minimum value of the data.
[0112] Calculating the real-time deviation of the data facilitates its subsequent use.
[0113] S234: The conversion is performed based on the ratio between the real-time deviation value and the baseline deviation value of the data.
[0114] Specifically, the data is normalized and quantified by calculating the ratio between the real-time deviation value and the baseline deviation value.
[0115] S235: When the conversion type is Boolean encoding, retrieve the category baseline data based on the parameter category, and perform the conversion based on the matching result between the category data and the category baseline data.
[0116] Among them, the category benchmark data refers to the preset standard data corresponding to the category of parameters.
[0117] When the conversion type is Boolean encoding, the category benchmark database is queried by parameter type to obtain category benchmark data. Then, the category data is normalized and quantized by matching the category data with the category benchmark data.
[0118] The category benchmark database pre-stores a table of different parameter categories and their corresponding category benchmark data, which is obtained after the operator pre-inputs the data.
[0119] For example, the category baseline database can be set to use idle data when the parameter category is vehicle status. When the parameter category is transport type, the corresponding category baseline data will be the data corresponding to special goods / dangerous goods.
[0120] S236: When the conversion type is hierarchical encoding type, retrieve the relevant type and related data based on the parameter type.
[0121] Among them, "related categories" refers to other parameter categories that are business-related to the parameter category to be converted and that will affect the accuracy of its quantification results. "Related data" refers to the specific information or values under the corresponding related category.
[0122] When the conversion type is hierarchical coding type, the relevant types are obtained by inputting the parameter type into the preset type-related database, and then the corresponding type data is retrieved based on the relevant types and used as relevant data for convenient subsequent use.
[0123] S237: Determine the correlation coefficient based on the relevant categories, and determine the relevant single reference value in combination with the relevant data.
[0124] The correlation coefficient refers to the degree of influence of each correlation category on the quantitative result of the current parameter category to be transformed. The single reference value of correlation refers to the single-dimensional quantitative reference value corresponding to the correlation data for each correlation category.
[0125] By determining the importance of the current parameter types and their business relationships based on historical scheduling data statistics or business rule presets, and then by quantifying and converting the corresponding relevant data according to the parameter types, a single reference value is obtained for convenient subsequent use.
[0126] S238: The relevant comprehensive reference value is obtained by weighting the correlation coefficient and the relevant single reference value.
[0127] In this method, the correlation coefficient and the relevant single reference value are weighted and the calculation result is used as the relevant comprehensive reference value for convenient subsequent use.
[0128] S239: Based on the parameter type, retrieve the reference level range and match it with the relevant comprehensive reference value for conversion.
[0129] The reference level range refers to the level range divided by score segments for parameter types. The reference level range must be stored in the system database based on the current business scheduling requirements for the parameter types, historical data distribution characteristics, and industry standards.
[0130] By matching the parameter types and retrieving the corresponding reference level ranges, and then matching the relevant comprehensive reference values with the reference level ranges, the categorical data is normalized and quantitatively transformed.
[0131] For example, when the parameter type is carrier assessment level, the corresponding related types are historical compliance rate and response time score. After retrieving the relevant data for historical compliance rate and quantifying it, the relevant single reference value is 0.85. After quantifying the relevant data for response time score, the relevant single reference value is 0.2. The correlation coefficient for historical compliance rate is 0.6, and the correlation coefficient for response time score is 0.5. At this point, the relevant comprehensive reference value is 0.61. Then, 0.61 is input into the reference level range for matching, thereby performing quantification.
[0132] S3: Determine the optimal matching scheduling strategy by combining historical scheduling data, task quantification values, and transportation capacity quantification values.
[0133] Among them, the optimal matching scheduling strategy refers to the initial scheduling scheme that can achieve precise matching between the current logistics task and transportation capacity resources.
[0134] By combining and analyzing historical scheduling data, task quantification values, and transportation capacity quantification values, the optimal matching scheduling strategy can be determined for subsequent use.
[0135] To further ensure the rationality of the optimal matching scheduling strategy, it is necessary to perform a further separate analysis and calculation on the optimal matching scheduling strategy, which will be explained in detail through the following steps.
[0136] The method for determining the optimal matching and scheduling strategy includes the following steps:
[0137] S31: Determine the basic weights and trigger condition set based on historical scheduling data.
[0138] Here, the basic weight refers to the initial impact weight of each task's quantified value and the transportation force's quantified value on the scheduling strategy matching result. The trigger condition set refers to the set of fluctuation judgment thresholds corresponding to each task's quantified value and the transportation force's quantified value.
[0139] By analyzing historical scheduling data, the basic weights and triggering condition set can be determined, which will facilitate subsequent use.
[0140] To further ensure the rationality of the basic weights and trigger condition set, it is necessary to conduct a further separate analysis and calculation of the basic weights and trigger condition set, which will be explained in detail through the steps shown below.
[0141] The method for determining the basic weights and the set of triggering conditions includes the following steps:
[0142] S311: Retrieve scheduling sample types and individual sample data based on historical scheduling data.
[0143] Here, "scheduling sample type" refers to the category to which each individual sample in the historical scheduling data belongs. "Individual sample data" refers to the data corresponding to each individual sample in the historical scheduling data. Historical scheduling data includes both scheduling sample types and individual sample data.
[0144] Historical scheduling data can be used to retrieve the types of scheduling samples and individual sample data for future use.
[0145] S312: Determine the maximum deviation value, median value, and reference value of the distribution based on a single sample of data.
[0146] Among them, the maximum deviation value of the distribution refers to the maximum difference between samples obtained after statistical analysis of single sample data of the same scheduling sample type. The median value of the distribution refers to the median value between samples obtained after statistical analysis of single sample data of the same scheduling sample type. The distribution degree reference value refers to the statistical value reflecting the overall dispersion and distribution characteristics of single sample data of the same scheduling sample type.
[0147] By performing statistical analysis on single sample data, selecting the median value as the distribution median, calculating the maximum difference between samples as the distribution maximum deviation value, and then calculating the dispersion of each sample from the distribution median value to obtain a distribution degree reference value for convenient subsequent use.
[0148] S313: Determine the trigger selection value based on the distribution level reference value.
[0149] The trigger selection value refers to the reference coefficient used to select the degree of deviation.
[0150] By inputting the distribution level reference value into the preset trigger selection database, a trigger selection value is obtained for convenient subsequent use.
[0151] The higher the distribution level reference value, the higher the trigger selection value. The trigger selection database pre-stores a lookup table of different distribution level reference values and their corresponding trigger selection values, and is obtained after the operator has pre-entered the data.
[0152] S314: Determine the number of categories based on the types of scheduled samples.
[0153] Among them, the number of categories refers to the number of categories corresponding to each scheduling sample category.
[0154] By counting the types of scheduled samples and using the count results as the number of types, it is convenient to use them later.
[0155] S315: Determine the single-category weight by combining the number of categories and the reference value of distribution degree.
[0156] Among them, single-category weight refers to the initial weight required for a single category.
[0157] By calculating the total effective sample size for all sample types, then calculating the proportion of each individual type, and finally weighting the types by weighting the number of types with reference distribution values, a single-type weight can be obtained for convenient subsequent use. The specific weights for the weighting calculation are preset by the operator according to actual needs.
[0158] S316: Calculate the product between the selected trigger value and the maximum distribution deviation value and use it as the single trigger deviation value for the distribution.
[0159] Among them, the single-trigger deviation value refers to the deviation value corresponding to the need for triggering adjustment.
[0160] The product between the selected trigger value and the maximum deviation value of the distribution is calculated, and the result is used as the single trigger deviation value of the distribution for convenient subsequent use.
[0161] S317: Combine the single weight of each type of scheduling sample with the single triggering deviation value of each type of scheduling sample as the basic weight, and combine the distribution median value and distribution single triggering deviation value of each type of scheduling sample as the triggering condition set.
[0162] Specifically, by combining the single weights corresponding to each type of scheduling sample, a basic weight is obtained. Then, by using the median value of the distribution and the single trigger deviation value corresponding to each type of scheduling sample as the two endpoints of the threshold range, the trigger tolerance range is obtained. Finally, the trigger tolerance ranges corresponding to each type of scheduling sample are combined to form the trigger condition set, thereby improving the accuracy of the obtained basic weights and trigger condition set.
[0163] S32: Based on the task quantification value, the transportation quantification value and the set of triggering conditions, the fluctuation adjustment value is determined by matching and comparing them.
[0164] Among them, the fluctuation adjustment value refers to the adjustment value corresponding to the adjustment of the basic weight when the quantified value deviates from the threshold range.
[0165] By matching and comparing the task quantification value, the transportation quantification value, and the set of triggering conditions, the fluctuation adjustment value is determined for convenient subsequent use.
[0166] To further ensure the rationality of the volatility adjustment value, it is necessary to perform a further separate analysis and calculation on the volatility adjustment value, which will be explained in detail through the steps shown below.
[0167] The method for determining the fluctuation adjustment value includes the following steps:
[0168] S321: Select the corresponding distribution median value from the set of triggering conditions based on the task quantification value or the transportation quantification value, and use it as the selection benchmark value.
[0169] The selection and definition of the reference value facilitates its subsequent use.
[0170] S322: Define the task quantification value or the transportation quantification value as the current selection value, and retrieve the current selection type based on the current selection value.
[0171] Among them, the currently selected category refers to the category to which the currently selected value belongs.
[0172] By defining the currently selected value and retrieving the currently selected type, it is convenient to use it later.
[0173] S323: Calculate the difference between the current selected value and the selected reference value and use it as the selection deviation value.
[0174] The selected deviation value refers to the difference between the current selected value and the selected reference value.
[0175] Calculating the selected deviation value facilitates subsequent use.
[0176] S324: Calculate the ratio between the selected deviation value and the selected reference value and use it as the selection ratio value.
[0177] The selected ratio value refers to the ratio between the selected deviation value and the selected benchmark value.
[0178] Calculating the selected ratio value facilitates subsequent use.
[0179] S325: Determine the species sensitivity coefficient based on the currently selected species.
[0180] Among them, the type sensitivity coefficient is a preset coefficient value used to indicate the degree of influence of the currently selected type on the matching effect of the scheduling strategy.
[0181] The currently selected category is entered into a preset sensitivity coefficient database to obtain the category sensitivity coefficient, which is convenient for subsequent use.
[0182] The sensitivity coefficient database pre-stores a lookup table of different currently selected categories and their corresponding sensitivity coefficients. The sensitivity coefficient database is obtained after the operator pre-inputs the data.
[0183] S326: Calculate the product between the selected proportion value and the type sensitivity coefficient and use it as the fluctuation adjustment value.
[0184] Specifically, the product between the selected proportion and the type sensitivity coefficient is calculated, and the result is used as a fluctuation adjustment value for convenient subsequent use.
[0185] S33: Adjust the base weights based on the fluctuation adjustment value to obtain the adjusted weights.
[0186] Specifically, the adjustment weight is obtained by calculating the sum of the fluctuation adjustment value and 1, and then calculating the product between the sum and the basic weight, which is convenient for subsequent use.
[0187] S34: Input the adjusted weights, task quantification values, and transportation quantification values into the preset integer programming model to obtain the planning and scheduling strategy, and use the planning and scheduling strategy as the optimal matching scheduling strategy.
[0188] Among them, the integer programming model is a pre-defined mathematical model specifically for logistics scheduling. It uses business objectives such as timeliness priority, cost control, and optimal resource allocation as constraints, supporting the optimal combination of discrete transportation resources and task requirements. The planning and scheduling strategy refers to the scheduling strategy corresponding to the planning based on adjusted weights.
[0189] By adjusting the weights, task quantification values, and transportation quantification values and inputting them into a preset integer programming model to obtain a planning and scheduling strategy, and using the planning and scheduling strategy as the optimal matching scheduling strategy, the accuracy of the obtained optimal matching scheduling strategy is improved.
[0190] S4: Dynamically optimize based on the optimal matching scheduling strategy to generate an optimized scheduling strategy.
[0191] Among them, the optimized scheduling strategy refers to the scheduling strategy corresponding to the optimized matching scheduling strategy.
[0192] The optimal matching scheduling strategy is dynamically optimized using a preset optimization algorithm to generate an optimized scheduling strategy, which is convenient for subsequent use.
[0193] To further ensure the rationality of the optimized scheduling strategy, it is necessary to conduct a further separate analysis and calculation of the optimized scheduling strategy, which will be explained in detail through the following steps.
[0194] The method for generating optimized scheduling strategies includes the following steps:
[0195] S41: Select the type of algorithm for data acquisition optimization.
[0196] The term "optimization selection algorithm type" refers to the type of optimization algorithm chosen when optimizing the optimal matching and scheduling strategy. Optimization selection algorithm types include improved genetic algorithms, NSGA-II algorithms, tabu search algorithms, and DQN reinforcement learning algorithms.
[0197] S42: Retrieve the numerical values of a single sample based on a single sample data.
[0198] Here, the number of samples refers to the total number of data points corresponding to a single sample.
[0199] By counting the number of samples in a single sample data set and using the count result as the sample count value, it is convenient to use later.
[0200] S43: Determine the algorithm configuration parameters by combining the optimization selection of algorithm types, the number of samples, and the number of types.
[0201] Among these, algorithm configuration parameters refer to the core adjustable parameters for each optimization algorithm type during runtime. For improved genetic algorithms, these parameters include the number of iterations, population size, crossover probability, and mutation probability. For NSGA-II algorithms, these parameters include crowding threshold and selection probability. For DQN reinforcement learning algorithms, these parameters include learning rate, discount factor, and experience replay pool size.
[0202] The average number of samples for a single category is determined by calculating the quotient between the number of samples and the number of categories. The average number of samples and the selected algorithm categories are then input into a preset algorithm configuration database to obtain the algorithm configuration parameters for subsequent use.
[0203] Different optimization selection algorithms and average sample sizes correspond to different algorithm configuration parameters. The algorithm configuration database pre-stores a table showing the correspondence between different optimization selection algorithms, average sample sizes, and corresponding algorithm configuration parameters. The algorithm configuration database is obtained after the operator pre-inputs the information.
[0204] S44: Configure the algorithm corresponding to the selected algorithm type based on the algorithm configuration parameters and use it as the optimization configuration algorithm.
[0205] Specifically, the algorithm is configured with algorithm configuration parameters to select the appropriate algorithm type for optimization, and the configured algorithm is used as the optimization configuration algorithm for convenient use in the future.
[0206] S45: The optimal matching scheduling strategy is input into the optimization configuration algorithm to obtain the optimized scheduling strategy.
[0207] Specifically, by inputting the optimal matching scheduling strategy into the optimization configuration algorithm, the optimization configuration algorithm optimizes the optimal matching scheduling strategy to obtain an optimized scheduling strategy, thereby improving the accuracy of the obtained optimized scheduling strategy.
[0208] S5: Perform logistics scheduling based on optimized scheduling strategies and collect parameters for abnormal situations.
[0209] Among them, the abnormal situation parameter refers to the data set of various unexpected abnormal events that occur during the logistics scheduling process. The abnormal situation parameter achieves real-time identification of anomalies through multi-source data acquisition (GPS, sensors, manual reporting, road condition / weather API).
[0210] By employing optimized scheduling strategies for logistics scheduling and collecting parameters for abnormal situations in real time, it is convenient for subsequent use.
[0211] S6: Determine the scheduling adjustment strategy by combining the abnormal situation parameters and the optimized scheduling strategy, and execute the scheduling adjustment strategy.
[0212] Among them, the scheduling adjustment strategy refers to the strategy corresponding to the adjustment of the scheduling strategy based on abnormal situations.
[0213] By combining and analyzing abnormal situation parameters with optimized scheduling strategies, scheduling adjustment strategies are determined and executed. This ensures that logistics scheduling not only conforms to historical scheduling experience but also responds to real-time tasks, capacity changes, and sudden anomalies, making it convenient for logistics scheduling to adapt to the full-dimensional needs of complex logistics scenarios.
[0214] To further ensure the rationality of the scheduling adjustment strategy, it is necessary to conduct a more detailed separate analysis and calculation of the scheduling adjustment strategy, which will be explained in detail through the steps shown below.
[0215] The method for determining the scheduling adjustment strategy includes the following steps:
[0216] S61: Retrieve the anomaly type and its impact range based on the anomaly parameters.
[0217] Among these, "abnormality type" refers to the category of abnormality in the abnormality parameters. "Affected scope" refers to the quantitative degree to which the abnormality impacts logistics scheduling. Abnormality types include vehicle malfunctions, traffic accidents, weather-related impacts, road closures, and cargo abnormalities. The abnormality parameters include both the abnormality type and its affected scope.
[0218] The abnormal situation parameters can be used to retrieve the abnormal type and its scope of impact, which is convenient for subsequent use.
[0219] S62: Determine the response level based on the anomaly type and the scope of impact.
[0220] The response level refers to the designated level of emergency response to an anomaly.
[0221] The response level is obtained by inputting the exception type and its impact range into a preset response level database, which facilitates subsequent use.
[0222] The response level database pre-stores a table showing the correspondence between different anomaly types, their impact ranges, and their corresponding response levels. The response level database is retrieved after the operator has pre-entered the information.
[0223] S63: Determine the adjustment type based on the response level.
[0224] Among them, adjustment type refers to the type corresponding to the adjustment of the strategy.
[0225] Different response levels correspond to different adjustment types. The higher the response level, the greater the magnitude and scope of the corresponding adjustment actions, and the more resources are invested in handling them.
[0226] S64: Determine the intervention and adjustment strategy by combining the adjustment type and the abnormality type.
[0227] Among them, the intervention and adjustment strategy refers to the scheduling strategy corresponding to the intervention.
[0228] By inputting the adjustment type and the abnormality type into a preset intervention adjustment database, an intervention adjustment strategy can be obtained for easy subsequent use.
[0229] The intervention adjustment database pre-stores a table of different adjustment types, abnormality types, and corresponding intervention adjustment strategies. The intervention adjustment database is obtained after the operator pre-inputs the data.
[0230] S65: Adjust the optimized scheduling strategy based on the intervention and adjustment strategy to obtain the scheduling adjustment strategy.
[0231] Specifically, by parsing the execution items between the intervention adjustment strategy and the optimization scheduling strategy, replacing them based on the same execution items, and then combining all the replaced execution items as the scheduling adjustment strategy, the accuracy of the obtained scheduling adjustment strategy is improved.
[0232] Based on the same inventive concept, embodiments of the present invention provide a multi-dimensional dynamic fusion intelligent logistics scheduling system, including:
[0233] The data acquisition module is used to collect logistics task parameters, transportation capacity resource parameters, historical scheduling data, abnormal situation parameters, and to select optimization algorithms.
[0234] The memory stores a program for implementing a multi-dimensional dynamic fusion intelligent logistics scheduling method as described above;
[0235] The processor loads and executes programs stored in memory.
[0236] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.
[0237] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.
Claims
1. A multi-dimensional dynamic fusion intelligent logistics scheduling method, characterized in that, include: Collect logistics task parameters, transportation capacity resource parameters, and historical scheduling data; The task parameters are quantified and converted to generate a task quantification value, and the transportation capacity resource parameters are quantified and converted to generate a transportation capacity quantification value. The optimal matching scheduling strategy is determined by combining historical scheduling data, task quantification values, and transportation capacity quantification values. Dynamically optimize the scheduling strategy based on the optimal matching strategy to generate an optimized scheduling strategy; Logistics scheduling is carried out based on optimized scheduling strategies, and parameters for abnormal situations are collected. The scheduling adjustment strategy is determined by combining abnormal situation parameters with the optimized scheduling strategy, and then the scheduling adjustment strategy is executed.
2. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 1, characterized in that, The specific steps of quantization conversion include: Retrieve parameter types and data based on logistics task parameters or transportation capacity resource parameters; The conversion type is determined based on the type of parameters. Conversion types include hierarchical encoding type, normalization type, and Boolean encoding type. Transform category data based on transformation type.
3. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 2, characterized in that, Transformation of categorical data based on transformation type includes: Retrieve the maximum, minimum, and real-time values of data based on the category data; When the conversion type is normalization, the difference between the maximum and minimum data values is calculated and used as the data baseline deviation value. Calculate the difference between the real-time data value and the minimum data value and use it as the real-time deviation value of the data; The conversion is performed based on the ratio between the real-time deviation value and the baseline deviation value of the data. When the conversion type is Boolean encoding, the category baseline data is retrieved based on the parameter category, and the conversion is performed based on the matching result between the category data and the category baseline data.
4. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 2, characterized in that, Transforming category data based on transformation type also includes: When the conversion type is hierarchical encoding, the relevant type and related data are retrieved based on the parameter type. The correlation coefficient is determined based on the relevant categories, and a single reference value is determined in combination with the relevant data; A comprehensive reference value is obtained by weighting the correlation coefficient with the relevant single reference value. The reference level range is retrieved based on the parameter type and matched with relevant comprehensive reference values for conversion.
5. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 1, characterized in that, Methods for determining the optimal matching and scheduling strategy include: Determine the basic weights and trigger condition set based on historical scheduling data; The fluctuation adjustment value is determined by matching and comparing the task quantification value, the transportation quantification value, and the set of triggering conditions. The base weights are adjusted based on the fluctuation adjustment value to obtain the adjusted weights; The adjusted weights, task quantification values, and transportation quantification values are input into a preset integer programming model to obtain a planning and scheduling strategy, which is then used as the optimal matching scheduling strategy.
6. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 5, characterized in that, The methods for determining the basic weights and trigger condition set include: Retrieve scheduling sample types and individual sample data based on historical scheduling data; Determine the maximum deviation value, median value, and reference value of the distribution based on single sample data; The trigger selection value is determined based on the distribution level reference value; The numerical value of each category is determined based on the category of the scheduling sample; The weight of a single category is determined by combining the numerical value of the number of categories with the reference value of the degree of distribution. Calculate the product of the selected trigger value and the maximum distribution deviation value, and use it as the single trigger deviation value of the distribution. The basic weight is the single weight corresponding to each type of scheduling sample, and the trigger condition set is the distribution median value and the single trigger deviation value corresponding to each type of scheduling sample.
7. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 6, characterized in that, Methods for determining fluctuation adjustment values include: Based on the task quantification value or the transportation force quantification value, select the corresponding distribution median value from the set of triggering conditions and use it as the selection benchmark value; Define a task quantification value or a transportation quantification value as the current selection value, and retrieve the current selection type based on the current selection value; Calculate the difference between the current selected value and the selected reference value and use it as the selection deviation value; Calculate the ratio between the selected deviation value and the selected benchmark value, and use it as the selection ratio value; Determine the sensitivity coefficient of the species based on the currently selected species; Calculate the product between the selected proportion value and the type sensitivity coefficient, and use it as the fluctuation adjustment value.
8. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 6, characterized in that, Methods for generating optimized scheduling strategies include: Selecting and optimizing algorithm types for data acquisition; Retrieve numerical values from a single sample of data; The algorithm configuration parameters are determined by combining the selection of algorithm types, the number of samples, and the number of types. Based on the algorithm configuration parameters, the algorithm corresponding to the selected algorithm type is configured and used as the optimization configuration algorithm; The optimal matching scheduling strategy is input into the optimization configuration algorithm to obtain the optimized scheduling strategy.
9. The multi-dimensional dynamic fusion intelligent logistics scheduling method according to claim 1, characterized in that, Methods for determining scheduling adjustment strategies include: Retrieve the anomaly type and its impact range based on the abnormal situation parameters; The response level is determined by combining the anomaly type and the scope of its impact; The adjustment type is determined based on the response level; The intervention and adjustment strategy is determined by combining the adjustment type and the abnormality type; The optimal scheduling strategy is adjusted based on the intervention and adjustment strategy, and the scheduling adjustment strategy is obtained.
10. A multi-dimensional dynamic fusion intelligent logistics scheduling system, characterized in that, include: The data acquisition module is used to collect logistics task parameters, transportation capacity resource parameters, historical scheduling data, abnormal situation parameters, and to select optimization algorithms. The memory stores a program for implementing a multi-dimensional dynamic fusion intelligent logistics scheduling method as described in any one of claims 1 to 9; The processor loads and executes programs stored in memory.