A supply and demand level determination method and device

Abstract

The specification discloses a supply-demand level determination method and device, which determines feature values of each feature dimension for representing a supply-demand state at a current time and classifications corresponding to each feature dimension, and then determines membership degrees of each feature dimension belonging to each classification. Then, the classifications of each feature dimension are arranged and combined to obtain each classification combination, and then the membership degrees of the supply-demand state at the current time belonging to each classification combination and an uncertainty measure are determined according to the membership degrees of each feature dimension belonging to each classification, and the supply-demand level corresponding to the supply-demand state at the current time is determined according to a preset corresponding relationship between each classification combination and the supply-demand level, so as to determine a service scheduling strategy according to the obtained supply-demand level. The supply-demand level corresponding to the current supply-demand state is determined according to the feature values of each feature dimension representing the current supply-demand state, so as to adjust the service scheduling strategy according to the current supply-demand level, thereby improving the robustness of service scheduling.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method and apparatus for determining supply and demand levels. Background Technology

[0002] Currently, with the development of on-demand delivery services, on-demand delivery orders have increased significantly. In order to improve the experience for all parties involved in delivery, it is necessary to make appropriate scheduling decisions and rationally dispatch delivery personnel to perform delivery tasks.

[0003] In existing technologies, various indicators are generally considered, and a multi-objective evaluation method with linear weighting is used to determine the evaluation score for each delivery person. Orders are then assigned to the delivery person with the best evaluation. For example, when assigning orders, the on-time delivery rate and the order's route convenience for each rider can be considered. These two indicators are weighted according to preset weights and summed to determine the evaluation score for each delivery person. The order is then assigned to the delivery person with the best evaluation.

[0004] However, in practice, the relationship between delivery capacity and order volume is related to the emphasis placed on various indicators during evaluation. For example, with constant delivery capacity, as order volume increases, it's necessary to ensure orders are delivered on time to guarantee user experience; in this case, the focus should be on the on-time delivery rate. However, when order volume is extremely high, it's necessary to improve delivery efficiency to ensure all orders are delivered; in this case, the focus should be on the orderliness of delivery routes.

[0005] Current multi-objective evaluation methods do not link supply and demand status with the degree of emphasis on various indicators, but only determine the weighted sum of each indicator by the same weight, resulting in low robustness of scheduling decisions. Summary of the Invention

[0006] This specification provides a method and apparatus for determining supply and demand levels, which at least partially solves the problems existing in the prior art.

[0007] The following technical solution is adopted in this specification:

[0008] This specification provides a method for determining supply and demand levels, including:

[0009] Determine the feature values ​​of each feature dimension used to characterize the supply and demand state at the current moment, and the corresponding classification for each feature dimension;

[0010] For each feature dimension, based on the category corresponding to that feature dimension, determine the membership degree of the feature value of that feature dimension to each category, and obtain the membership degree vector corresponding to that feature dimension.

[0011] The classification of each feature dimension is arranged and combined to obtain each classification combination. Based on each classification combination and the membership vector corresponding to each feature dimension, the membership degree and uncertainty measure of the supply and demand state at the current moment to each classification combination are determined.

[0012] Based on the preset correspondence between each category combination and the supply and demand level, the membership degree of the current supply and demand status to each category combination, and the uncertainty measure, the supply and demand level corresponding to the current supply and demand status is determined, so as to determine the business scheduling strategy based on the obtained supply and demand level.

[0013] Optionally, the classification corresponding to each feature dimension is determined, specifically including:

[0014] For each feature dimension, when the feature dimension is a continuous variable, obtain the historical feature value of the feature dimension in the preset historical number of days, and sort the historical feature values.

[0015] Based on the preset percentiles and the sorting, the historical feature values ​​corresponding to each percentile index are determined and used as the classification threshold corresponding to that feature dimension.

[0016] Based on the classification threshold corresponding to the feature dimension, determine each classification threshold interval corresponding to the feature dimension, wherein at least some of the classification threshold intervals overlap.

[0017] The category corresponding to the feature dimension is determined based on the classification threshold ranges corresponding to each feature dimension.

[0018] When the feature dimension is a discrete variable, the category corresponding to the feature dimension is determined according to the preset classification threshold.

[0019] Optionally, based on the classification corresponding to the feature dimension, the membership degree of each feature value of the feature dimension to each classification is determined, resulting in the membership vector corresponding to the feature dimension, specifically including:

[0020] Based on the classification corresponding to this feature dimension, determine the membership degree of the feature value of this feature dimension to each classification, and use it as the coarse classification result;

[0021] Determine whether the feature dimension is a preset feature dimension that needs to be further classified;

[0022] If so, then based on at least some of the sub-classification thresholds corresponding to the feature dimension, the feature dimension is reclassified, and based on the classification determined after the sub-classification of the feature dimension, the membership degree of the feature value of the feature dimension to each classification is determined, and the membership degree vector corresponding to the feature dimension is obtained.

[0023] If not, then determine the membership vector corresponding to that feature dimension based on the coarse classification result.

[0024] Optionally, the classifications of each feature dimension can be permuted and combined to obtain various classification combinations, specifically including:

[0025] For each feature dimension, the membership vector corresponding to the membership degree is determined, and the classification corresponding to the membership degree that meets the preset conditions is used as the effective classification of that feature dimension.

[0026] The effective classifications of each feature dimension are arranged and combined to obtain various classification combinations.

[0027] Optionally, based on the membership vectors corresponding to each classification combination and each feature dimension, the membership degree and uncertainty measure of the supply and demand state at the current moment belonging to each classification combination are determined, specifically including:

[0028] For each classification combination, the membership degree of each feature value in the classification combination is determined based on the membership vector corresponding to each feature dimension.

[0029] The membership degree of the supply and demand status at the current moment to the corresponding category is determined by the product of the membership degrees of the feature values ​​of each feature dimension to the corresponding category in the category combination.

[0030] The uncertainty measure of the supply and demand state belonging to the category combination at the current moment is determined by the ratio of the first cumulative sum of the differences between the membership degree of the supply and demand state at the current moment belonging to each other category combination and the membership degree of the supply and demand state at the current moment belonging to that category combination, to the second cumulative sum of the membership degree of the supply and demand state at the current moment belonging to each category combination.

[0031] Optionally, the supply and demand level corresponding to the current supply and demand state is determined based on the preset correspondence between each category combination and the supply and demand level, the membership degree of the current supply and demand state to each category combination, and the uncertainty measure. Specifically, this includes:

[0032] Based on the product of the membership degree and uncertainty measure of the supply and demand status at the current moment to each category combination, determine the category combination corresponding to the product when the product reaches its maximum.

[0033] Based on the classification combination and the preset correspondence between each classification combination and the supply and demand level, the supply and demand level corresponding to the classification combination is determined as the supply and demand level corresponding to the current supply and demand status.

[0034] Optionally, a business scheduling strategy is determined based on the obtained supply and demand levels, specifically including:

[0035] Based on the acquired movement information and order information of each user, as well as the order information to be assigned, the delivery route for each user to deliver the order to be assigned is determined;

[0036] For each user, based on the user's delivery route, determine the probability distribution of the completion time of each order, and sample the completion time of each order to determine the order timeout situation corresponding to each sample;

[0037] Based on the order timeout status corresponding to each sample and the preset timeout evaluation function, determine the expected timeout evaluation and conditional risk timeout evaluation for the user to complete delivery;

[0038] Based on the obtained supply and demand levels, the relationship between the preset supply and demand levels and the first weight corresponding to the expected timeout evaluation, and the relationship between the supply and demand levels and the second weight corresponding to the conditional risk timeout evaluation, the first weight and the second weight are determined.

[0039] The expected timeout evaluation and the conditional risk timeout evaluation corresponding to each user are weighted and summed according to the first weight and the second weight, and the orders to be allocated are assigned to the user corresponding to the one whose weighted sum is minimized.

[0040] This specification provides a supply and demand level determination device, including:

[0041] The classification module is used to determine the feature values ​​of each feature dimension used to represent the supply and demand state at the current moment, as well as the corresponding classification of each feature dimension;

[0042] The membership determination module is used to determine the membership degree of each feature dimension to each category based on the category corresponding to that feature dimension, and obtain the membership degree vector corresponding to that feature dimension.

[0043] The classification combination determination module is used to arrange and combine the classifications of each feature dimension to obtain each classification combination. Based on each classification combination and the membership vector corresponding to each feature dimension, the module determines the membership degree and uncertainty measure of the supply and demand state at the current moment to each classification combination.

[0044] The supply and demand level determination module is used to determine the supply and demand level corresponding to the current supply and demand state based on the preset correspondence between each category combination and the supply and demand level, the membership degree of the current supply and demand state to each category combination, and the uncertainty measure, so as to determine the business scheduling strategy based on the obtained supply and demand level.

[0045] This specification provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described supply and demand level determination method.

[0046] This specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described supply and demand level determination method.

[0047] The above-mentioned technical solutions adopted in this specification can achieve the following beneficial effects:

[0048] The supply and demand level determination method provided in this specification first determines the feature values ​​of each feature dimension representing the supply and demand state at the current moment, as well as the corresponding classification for each feature dimension. Then, it determines the membership degree of each feature dimension to each classification. Next, it arranges and combines the classifications of each feature dimension to obtain various classification combinations. Then, based on the membership degrees of each feature dimension to each classification, it determines the membership degree and uncertainty measure of the current supply and demand state to each classification combination. Finally, based on the preset correspondence between each classification combination and the supply and demand level, it determines the supply and demand level corresponding to the current supply and demand state, and uses this obtained supply and demand level to determine the business scheduling strategy. By determining the supply and demand level corresponding to the current supply and demand state based on the feature values ​​of each feature dimension representing the current supply and demand state, and adjusting the business scheduling strategy according to the current supply and demand level, the robustness of business scheduling is improved. Attached Figure Description

[0049] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0050] Figure 1 A flowchart illustrating a method for determining supply and demand levels provided in this specification;

[0051] Figure 2 This is a schematic diagram of a membership function provided in this specification;

[0052] Figure 3 This is a schematic diagram illustrating the correspondence between various classification combinations and supply and demand levels provided in this specification;

[0053] Figure 4 This is a schematic diagram of a membership function after fine classification, provided in this specification.

[0054] Figure 5 This is a schematic diagram of a supply and demand level determination device provided in this specification;

[0055] Figure 6 This is a schematic diagram of an electronic device for implementing a method for determining supply and demand levels, as provided in this specification. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without creative effort are within the scope of protection of this application.

[0057] Currently, in the on-demand delivery sector, order scheduling strategies typically involve a combination of various evaluation metrics, using a linearly weighted multi-objective evaluation method to allocate orders to the best-rated delivery personnel. However, in reality, the supply and demand situation should be correlated with the degree of emphasis placed on different metrics during the evaluation process.

[0058] For example, in a given region, with total delivery capacity remaining constant, as order volume increases, it's crucial to ensure timely delivery to guarantee a positive user experience. In this case, the on-time delivery rate of delivery personnel should be the primary consideration in decision-making. However, when order volume is extremely high, it's necessary to improve delivery efficiency to alleviate delivery pressure and ensure all orders are completed. In this situation, the convenience of delivery routes should be the primary consideration in decision-making.

[0059] However, the current multi-objective evaluation method does not link the supply and demand status with the degree of emphasis on various indicators. Even in different supply and demand states, various evaluation indicators are still linearly weighted with the same weight to determine the evaluation of each deliveryman and thus determine the dispatch strategy, resulting in low adaptability and robustness of the dispatch strategy.

[0060] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0061] Figure 1 This is a flowchart illustrating one method for determining supply and demand levels in this specification, specifically including the following steps:

[0062] S100: Determine the feature values ​​of each feature dimension used to characterize the supply and demand state at the current moment, as well as the corresponding classification of each feature dimension.

[0063] Currently, in the on-demand delivery sector, delivery personnel typically perform delivery tasks within pre-defined areas. Therefore, for each area, the business platform's server can periodically collect feature values ​​corresponding to the feature dimensions associated with the supply and demand status within that area. By analyzing these features, the supply and demand level corresponding to the current supply and demand status of that area can be determined, allowing for timely adjustments to order allocation strategies based on the current supply and demand situation.

[0064] Based on this, in one or more embodiments of this specification, the server of the business platform can determine the feature values ​​of each feature dimension used to characterize the supply and demand status of a certain region at the current time, as well as the classification threshold corresponding to each feature dimension, according to a preset time interval.

[0065] The specific feature dimensions to be collected can be determined as needed, and this manual does not impose any restrictions on this. For example, it could be the average number of orders received by each delivery person in the region (referred to as load, x). load The number of orders generated in this area within the 10 minutes prior to the data collection time (referred to as the order volume, x) wb ), the weather in the area at the time of collection (x wth )etc.

[0066] When determining the category corresponding to each feature dimension, the server of the business platform can first obtain the historical feature values ​​of the feature dimension in the preset historical number of days for each feature dimension, and sort the historical feature values.

[0067] Then, based on the preset percentiles and the above sorting, the historical feature values ​​corresponding to each percentile index are determined as the classification thresholds corresponding to that feature dimension.

[0068] Secondly, based on the classification threshold corresponding to the feature dimension, determine the classification threshold intervals corresponding to the feature dimension, with at least some of the classification threshold intervals overlapping.

[0069] Finally, the category corresponding to the feature dimension is determined based on the classification threshold ranges corresponding to each feature dimension.

[0070] In this context, a continuous variable feature dimension means that its feature values ​​change continuously, such as the load and order volume mentioned above. The percentile metric refers to the value that corresponds to the percentile position among all data after sorting. At least partial overlap of classification threshold intervals is used to perform fuzzy classification of the feature values ​​of the feature dimension, thereby avoiding similar feature values ​​at threshold boundaries being assigned to different categories. Of course, if the feature dimension is a continuous variable and the corresponding classification threshold intervals do not overlap, then it follows the same principle as if the feature dimension were a discrete variable.

[0071] For example, assuming a preset historical period of 10 days, the business platform's server collects feature values ​​for each characteristic dimension every 10 minutes, with preset percentiles of 50%, 80%, and 95%, then the percentile indicators are TP50, TP80, and TP95. Taking load as an example, 1440 historical feature values ​​for the load characteristic dimension over the past 10 days can be obtained. These 1440 data points are sorted in ascending order. The 720th historical feature value (referred to as load) is then... 50 ), the 1152nd historical feature (abbreviated as load) 80 ) and the 1368th historical feature (referred to as load) 95 This refers to the preset historical classification threshold corresponding to the load. Then, based on the determined classification threshold, the low load threshold range can be determined as [0, load...]. 80 The load threshold range is [load] 50 load 95 The high load threshold range is [load] 80 [+∞]. It is clear that the category corresponding to this load characteristic dimension is low load, medium load, and high load.

[0072] For features that are discrete variables, the business platform's server can determine the category corresponding to that feature dimension based on a preset classification threshold. For example, weather can be generally evaluated: a preset classification threshold of 10 indicates good weather, 20 indicates moderate weather, and 30 indicates severe weather. Clearly, the category corresponding to this weather feature dimension can be determined as good weather, moderate weather, or severe weather.

[0073] The server mentioned in this manual can refer to a server set up on a business platform, or a device such as a desktop computer or laptop computer capable of executing the solutions described in this manual. For ease of explanation, the following description will only focus on the server as the execution subject.

[0074] S102: For each feature dimension, based on the category corresponding to that feature dimension, determine the membership degree of the feature value of that feature dimension to each category, and obtain the membership degree vector corresponding to that feature dimension.

[0075] After obtaining the feature values ​​of each feature dimension representing the supply and demand state at the current moment, as well as the corresponding classification of each feature dimension, the server can determine the membership degree of each feature dimension's feature value to each classification. Subsequently, it can further determine the membership degree of different classification combinations corresponding to the current supply and demand state.

[0076] Specifically, in one or more embodiments of this specification, the server can determine the membership degree of each feature dimension to each category based on the category corresponding to the feature dimension and a preset membership function, thereby obtaining the membership degree vector corresponding to the feature dimension.

[0077] The membership function can be any function that maps eigenvalues ​​to the unit real number interval [0, 1], such as the triangular membership function, the trapezoidal membership function, etc. This specification does not restrict the specific type of membership function.

[0078] For example, continuing with the load feature dimension in step S100, assuming the server determines the load feature dimension to correspond to low load, medium load, and high load, the membership functions corresponding to different categories are as follows: Figure 2 As shown.

[0079] Figure 2 This is a schematic diagram of a membership function provided in this specification. Figure 2 Taking load as an example, this paper demonstrates how the load feature dimension is divided into three categories based on three determined classification thresholds: the thin solid line represents the membership function for low load, the thicker solid line represents the membership function for medium load, and the thickest solid line represents the membership function for high load. The membership functions corresponding to each category are as follows:

[0080]

[0081]

[0082]

[0083] The server can determine the membership degree of the feature values ​​of the load feature dimension to low load, medium load and high load respectively according to the membership function of each category, and use the vector composed of these three membership degrees as the membership degree vector corresponding to the load feature dimension.

[0084] For features where the dimension is a discrete variable, continuing with the weather feature dimension in step S100 as an example, the server can determine the evaluation of the weather at the current collection time based on a pre-set evaluation method. Therefore, the membership function corresponding to the "good weather" category is: The membership functions for other categories are similar and will not be elaborated here. The server can determine the membership degree of the feature values ​​of the weather feature dimension to good weather, normal weather, and severe weather based on the membership functions of each category, and use the vector composed of these three membership degrees as the membership degree vector corresponding to the weather feature dimension.

[0085] S104: Arrange and combine the classifications of each feature dimension to obtain each classification combination. Based on each classification combination and the membership vector corresponding to each feature dimension, determine the membership degree and uncertainty measure of the supply and demand state at the current moment to each classification combination.

[0086] After obtaining the classification corresponding to each feature dimension through step S100 and the membership vector corresponding to each feature dimension through step S102, in one or more embodiments of this specification, the server can first arrange and combine the classifications of each feature dimension to obtain each category combination, and then determine the membership degree and uncertainty measure of the supply and demand status at the current moment to each category combination based on the membership vectors corresponding to each feature dimension.

[0087] The classification combination mentioned here is obtained by arranging and combining the classifications of each feature dimension. That is, one classification is taken from the classifications of each feature dimension and then combined to obtain the combined classification.

[0088] For example, taking a feature dimension that includes both load and weather as an example, the classifications corresponding to each feature dimension can be found in the relevant content and will not be repeated here. After permuting and combining the classifications of each feature dimension, we can obtain nine classification combinations: (low load, good weather), (low load, normal weather), (low load, severe weather), (medium load, good weather), (medium load, normal weather), (medium load, severe weather), (high load, good weather), (high load, normal weather), and (high load, severe weather).

[0089] After determining each category combination, the server can further determine the membership degree and uncertainty measure of the current supply and demand status to each category combination.

[0090] Specifically, in one or more embodiments of this specification, for the membership degree of the current supply and demand status to the category combination, the server can first determine the membership degree of the feature value of each feature dimension to the corresponding category in the category combination based on the membership vector corresponding to each feature dimension for each category combination.

[0091] Then, based on the product of the membership degrees of the feature values ​​of each feature dimension to the corresponding category in the category combination, the membership degree of the supply and demand status at the current moment to the category combination is determined.

[0092] The membership degree of the current supply and demand status to each category combination can be calculated using the following formula:

[0093] u i，j，k，... (X)=u i (x1)×u j (x2)×u k (x3)×...

[0094] sti,j,k,...=1,2,3,...

[0095] In the formula, X represents the supply and demand state at the current moment, and x1, x2, and x3 represent the feature values ​​of each feature dimension determined in step S100. For ease of description, only three feature values ​​are specifically given here, and the remaining feature values ​​are omitted. i represents the ordinal number of a category for feature value x1, and the same applies to j, k, etc. u. i (x1) represents the membership degree of feature value x1 in the classification of feature dimension i, and the subsequent u j (x2), u k (x3) and so on are analogous. i, j, k, ... represent a combination of categories, such as category i, category j, etc. i，j，k，.. (X) represents the membership degree of the current supply and demand status to the category combination i, j, k, ...

[0096] For example, taking the above category combination (low load, good weather) as an example, let the category number of low load be 1 and the category number of good weather be 1, then u 1，1 (X)=u1(x load )×u1(x wth )=u(x load (low load) × u(x) wth Good weather).

[0097] Of course, for cases where the feature dimension is a discrete variable, the server can directly determine that the current supply and demand state belongs to a specific category based on the category corresponding to that feature dimension. For example, if the weather feature dimension is a discrete variable, the server can directly determine that the current supply and demand state belongs to a specific weather category based on the feature value of the weather feature dimension, while the membership degree to other weather categories will all be zero. Therefore, when determining category combinations and further determining the membership degree of the current supply and demand state to each category combination, for cases where the feature dimension is a discrete variable, the server can only determine the membership degree of the current supply and demand state to the category combinations obtained by combining categories whose feature dimension is not zero.

[0098] For the uncertain measure of whether the supply and demand status at the current moment belongs to the category combination, the server can determine the uncertain measure of whether the supply and demand status at the current moment belongs to the category combination based on the ratio of the first sum of the differences between the membership degree of the supply and demand status at the current moment belonging to each other category combination and the membership degree of the supply and demand status at the current moment belonging to the category combination, to the second sum of the membership degree of the supply and demand status at the current moment belonging to each category combination.

[0099] The uncertainty measure of whether the current supply and demand situation belongs to this category combination can be calculated using the following formula:

[0100]

[0101] sti,j,k,...=1,2,3,...

[0102] In the formula, P i，j，k，... (X) represents the uncertainty measure of the supply and demand state at the current moment belonging to the category combination i, j, k, ... . ω∈Ω\{i, j, k, ...} represents all other category combinations besides category combinations i, j, k, ... Therefore, u ω (X) represents the membership degree of the current supply and demand status to each of the other category combinations, ∑ ω∈Ω\{i，j，k，...} |u ω (X)-u i，j，k，... (X)| represents the first sum of the absolute values ​​of the differences between the membership degrees of the current supply and demand state to each of the other classification combinations and the membership degrees of the current supply and demand state to classification combinations i, j, k, ... . ∑u i，j，k，... (X) represents the second cumulative sum.

[0103] For cases where the feature dimension is a discrete variable, similar to membership, the server can determine only the uncertainty measure of the classification combination obtained by combining categories whose current supply and demand status belongs to categories whose feature dimension is not zero.

[0104] S106: Based on the preset correspondence between each category combination and the supply and demand level, the membership degree of the current supply and demand status to each category combination, and the uncertainty measure, determine the supply and demand level corresponding to the current supply and demand status, so as to determine the business scheduling strategy based on the obtained supply and demand level.

[0105] After obtaining the various category combinations and the membership degree and uncertainty measure of the current supply and demand status to each category combination, the server can determine the supply and demand level corresponding to the current supply and demand status based on the preset correspondence between each category combination and the supply and demand level, and then determine the business scheduling strategy based on the obtained supply and demand level.

[0106] Specifically, in one or more embodiments of this specification, the server can first determine the category combination corresponding to the product of the membership degree and the uncertainty measure of the current supply and demand status belonging to each category combination, and then determine the supply and demand level corresponding to the category combination as the supply and demand level corresponding to the current supply and demand status based on the preset correspondence between each category combination and the supply and demand level.

[0107] The correspondence between each category combination and the supply and demand level can be set as needed, such as... Figure 3 As shown.

[0108] Figure 3 This is a schematic diagram illustrating the correspondence between various classification combinations and supply and demand levels provided in this specification. Figure 3 Taking order volume, load, and weather as examples, each feature dimension corresponds to three categories. Order volume corresponds to low, medium, and high order volume categories, while the categories for load and weather can be found in the descriptions in step S100. Therefore, after permuting and combining these three feature dimensions in step S104, 27 category combinations can be obtained, i.e. Figure 3 The diagram contains 27 small boxes. For ease of description, these 27 small boxes are divided into three main categories based on weather conditions, each containing nine small boxes. Each row of a large box corresponds to a category within the order volume dimension, and each column corresponds to a category within the load dimension. The numbers within each small box represent the supply and demand level corresponding to that category. For example, the category combination represented by the thick solid-line small box is (low load, medium order volume, normal weather), corresponding to a supply and demand level of 2.

[0109] After obtaining the membership degree and uncertainty measure of the current supply and demand state to each category combination in step S104, the server can determine which category combination the current supply and demand state ultimately belongs to according to the following formula:

[0110] Y = argmax(u i，j，k，... (X)×P i，j，k，... (X))

[0111] sti,j,k,...=1,2,3,...

[0112] In the formula, u i，j，k，... (X) represents the membership degree of the current supply and demand status to each category combination, P i，j，k，... (X) represents the corresponding uncertainty measure. argmax(u i，j，k，... (X)×P i，j，k，... (X) represents the value of (i, g, k, ...) when the product of the membership degree and uncertainty measure of the supply and demand state at the current moment reaches its maximum, that is, the category combination corresponding to the maximum product.

[0113] For example, assuming the membership vector corresponding to the weather feature dimension is [0, 1, 0], step S104 can determine the supply and demand status at the current moment. Figure 3 The membership degree and uncertainty measure of the nine classification combinations corresponding to general weather conditions are determined. For each classification combination, the product of the membership degree and uncertainty measure corresponding to that combination is determined, and then the classification combination corresponding to the maximum value among the nine products is determined.

[0114] Then, based on the preset correspondence between each category combination and the supply and demand level, the supply and demand level corresponding to the obtained category combination can be determined, which is the supply and demand level corresponding to the current supply and demand status.

[0115] based on Figure 1 The method for determining supply and demand levels, as shown, first determines the feature values ​​of each feature dimension used to characterize the supply and demand state at the current moment, as well as the corresponding classification for each feature dimension. Then, it determines the membership degree of each feature dimension to its corresponding classification, obtaining a membership degree vector for each feature dimension. Next, it arranges and combines the classifications of each feature dimension to obtain various classification combinations. Based on the membership degree vectors corresponding to each feature dimension, it determines the membership degree and uncertainty measure of the current supply and demand state to each classification combination. Finally, based on the preset correspondence between each classification combination and the supply and demand level, the membership degree and uncertainty measure of the current supply and demand state to each classification combination, it determines the supply and demand level corresponding to the current supply and demand state, and uses this obtained supply and demand level to determine the business scheduling strategy. (Passed.)

[0116] Furthermore, in one or more embodiments of this specification, in step S102, in order to more accurately determine the supply and demand level corresponding to the supply and demand status at the current moment, when the server determines the membership degree of the feature value of each feature dimension to each category according to the category corresponding to each feature dimension, and obtains the membership degree vector corresponding to the feature dimension, it can also perform fine classification according to at least some feature dimensions.

[0117] Specifically, the server can first determine the membership degree of the feature value of the feature dimension to each category based on the category corresponding to the feature dimension, and use this as the coarse classification result.

[0118] Then, determine whether the feature dimension is a preset feature dimension that needs to be further classified.

[0119] If so, then based on at least some of the sub-classification thresholds corresponding to the feature dimension, the feature dimension is reclassified. Based on the classification determined after the sub-classification of the feature dimension, the membership degree of the feature value of the feature dimension to each classification is determined, and the membership degree vector corresponding to the feature dimension is obtained.

[0120] If not, then determine the membership vector corresponding to that feature dimension based on the coarse classification result.

[0121] The content of the coarse classification can be referred to the corresponding explanation in step S102, and will not be repeated here. The feature dimensions requiring fine classification can be preset as needed; this manual does not impose any restrictions on this. Of course, for features that are discrete variables, the membership degree of the feature dimension to each category is not fuzzy or uncertain; therefore, fine classification is usually not performed based on the feature dimensions of the corresponding discrete variables. The reclassification of the feature dimension based on at least some of the fine classification thresholds can be understood as determining the newly added fine classification threshold range based on at least some of the fine classification thresholds corresponding to the feature dimension, and then determining the newly added fine classification for the feature dimension based on the determined fine classification threshold range. The subsequent determination of the membership vector corresponding to the feature dimension can be referred to the corresponding explanation in step S102. As long as the membership degree of the feature dimension to each category is determined based on all categories corresponding to the feature dimension, the membership vector corresponding to the feature dimension can be obtained. The determination of the membership vector corresponding to the feature dimension based on the coarse classification results can be referred to the corresponding explanation in step S102.

[0122] For example, taking the load feature dimension in step S102 as an example, assuming the server determines that it needs to be further classified according to the load feature dimension, then the server can determine the low-to-medium load sub-classification threshold range as [load] based on the determined classification threshold. 50 load 80 The threshold range for medium to high load subcategories is [load]. 80 load 95 Clearly, the subcategories corresponding to this load characteristic dimension can be determined as low-to-medium load and medium-to-high load. Then, the server can determine the membership degree of each category for the characteristic value of this load characteristic dimension based on the preset membership functions for each category, thus obtaining the membership degree vector corresponding to this load characteristic dimension.

[0123] This specification provides a schematic diagram of membership functions after fine classification. For example... Figure 4 As shown.

[0124] Figure 4 In, with Figure 2 For the same parts, please refer to Figure 2 The corresponding explanations will not be repeated here. Compared to Figure 2 The two additional dashed lines represent the membership functions for the two subcategories—low-medium load and medium-high load—obtained after further subcategorization of the load feature dimensions. Specifically, the dotted dashed line represents the membership function for the low-medium load category, while the alternating dashed line segment and dot represents the membership function for the medium-high load category. The membership functions for the low-medium load and medium-high load categories are as follows:

[0125]

[0126]

[0127] By combining the membership functions of each category corresponding to the load feature dimension in step S102, the server can determine the membership degree of the feature value of the load feature dimension to low load, medium-low load, medium load, medium-high load, and high load, respectively, and use the vector composed of these 5 membership degrees as the membership degree vector corresponding to the load feature dimension.

[0128] Furthermore, in one or more embodiments of this specification, in step S104, corresponding to the above-mentioned detailed classification based on at least some feature dimensions, considering that there may be many classifications corresponding to each feature dimension, the server will also generate more classification combinations by arranging and combining the classifications of each feature dimension. In order to reduce the amount of computation, when determining the classification combination, the server may first determine the classifications corresponding to the membership vectors that meet the preset conditions for each feature dimension, and use these as the valid classifications for that feature dimension. Then, the valid classifications of each feature dimension are arranged and combined to obtain the classification combinations.

[0129] Among them, a classification that meets the preset conditions can be one whose membership degree is greater than a preset threshold. By arranging and combining only the effective classifications of each feature dimension, the classification combination is obtained, which greatly reduces the amount of subsequent calculation and saves computing resources.

[0130] For example, continuing with the aforementioned feature dimension of load, let's assume load 50 <x load <load 80 The preset condition is that the membership degree of this category is greater than zero. Then, based on the membership degree vector corresponding to each category in the load feature dimension, u(x) can be obtained. load High load) = 0, u(x) load (Medium-high load) = 0. The membership degrees of the feature values ​​of the load characteristic dimension belonging to the other three categories are all greater than zero. Therefore, the server can be classified as low load, medium-low load, and medium load as valid categories. For the content of the permutations and combinations to obtain each category combination, please refer to the corresponding explanation in step S104, which will not be repeated here.

[0131] In addition, in one or more embodiments of this specification, in step S106, the server may specifically determine the service scheduling strategy based on the obtained supply and demand levels through the steps described later.

[0132] Step 1: The server determines the delivery route for each user to deliver the order to be assigned based on the obtained movement information and order information of each user, as well as the order information to be assigned.

[0133] Step 2: For each user, the server determines the probability distribution of the completion time of each order based on the user's delivery route, and samples the completion time of each order several times to determine the order timeout situation corresponding to each sample.

[0134] When sampling the completion time of each order, the specific sampling method can be determined as needed, and this specification does not impose any restrictions. For example, assuming the number of samples is 10, and for a certain order, assuming that the completion time of the order conforms to a Gaussian distribution, the sampling interval can be the distribution interval with a probability greater than 5%, and then sampling can be performed from within the sampling interval using a uniform sampling method.

[0135] Step 3: The server determines the expected timeout rating and conditional risk timeout rating for the user to complete delivery based on the order timeout status corresponding to each sample and the preset timeout rating function.

[0136] The timeout evaluation function can be determined as needed, and this specification does not impose any restrictions on it. The expected timeout evaluation represents the mean of the timeout evaluations for each sample. The conditional risk timeout evaluation represents the conditional value at risk (CVaR) of the timeout evaluations for each sample.

[0137] Continuing with the above sampling example, suppose a user has a total of 3 orders, including those yet to be assigned. If each order is sampled 10 times, then the timeout minutes for the sample set are x = {x1, x2, ..., x...} 10}. Where x1 represents the sum of timeout cases corresponding to the first sampling of each of the three orders, x2 represents the sum of timeout cases corresponding to the second sampling of each of the three orders, and so on.

[0138] The expected timeout evaluation can be calculated using the following formula:

[0139]

[0140] In the formula, h e (x), E n [h(x n [)] represents the expected timeout evaluation. N represents the sample size, h(x) n ) represents the timeout evaluation function.

[0141] The following formula can be used to calculate the conditional risk timeout assessment:

[0142]

[0143] In the formula, h r (x), CVaR 1-β [h(x n[] represents conditional risk timeout assessment. As the sampling process shows, for each order, the larger the number of samples, the more severe the timeout situation. Conditional risk timeout assessment calculates the mean timeout assessment of the β% of samples with the most severe timeout assessments, i.e., The meaning of .

[0144] Of course, steps 1 to 3 are already mature technologies, and this manual will not go into detail about them.

[0145] Step 4: The server determines the first weight and the second weight based on the obtained supply and demand level, the relationship between the preset supply and demand level and the first weight corresponding to the expected timeout evaluation, and the relationship between the supply and demand level and the second weight corresponding to the conditional risk timeout evaluation.

[0146] The supply and demand levels can be determined through steps S100 to S106 above. The relationship between the supply and demand levels and the first weight can be set as needed, and this specification does not impose any restrictions on it.

[0147] For example, this specification provides a relationship between supply and demand levels and the first weight, as shown below:

[0148]

[0149] In the formula, α represents the first weight, and x represents the supply and demand level corresponding to the current supply and demand status. When the obtained supply and demand level is 3, the server can determine the first weight as 0.4.

[0150] Similarly, the server can determine the second weight based on the obtained supply and demand levels and the relationship between the preset supply and demand levels and the second weight, which will not be elaborated here.

[0151] Step 5: Based on the first weight and the second weight, perform a weighted sum on the expected timeout evaluation and the conditional risk timeout evaluation for each user, and allocate the order to be assigned to the user whose weighted sum is minimized.

[0152] After determining the expected timeout rating and conditional risk timeout rating for each user in step 3, and determining the first weight and second weight in step 4, the server can perform a weighted summation of the expected timeout rating and conditional risk timeout rating for each user based on the first weight and the second weight. When the weighted sum is minimized, it means that the user corresponding to the minimum weighted sum has the best rating, and the server can assign the order to be allocated to that user.

[0153] The weighted sum can be calculated using the following formula:

[0154] S=αh e (x)+γh r(x)

[0155] In the formula, S represents the weighted sum, α represents the first weight, and h r (x) represents the expected timeout evaluation, γ represents the second weight, and h r (x) indicates that the conditional risk timeout assessment has been completed.

[0156] Of course, steps 1 to 5 above are merely an example provided in this manual of determining a business scheduling strategy based on the obtained supply and demand levels. This manual does not limit the specific method for determining the business scheduling strategy based on the obtained supply and demand levels. For example, in step 4, the server can also determine the third weight and the fourth weight based on the obtained supply and demand levels, the relationship between the preset supply and demand levels and the third weight corresponding to the expected time consumption evaluation, and the relationship between the supply and demand levels and the fourth weight corresponding to the conditional risk time consumption evaluation. Then, in the subsequent step 5, the corresponding evaluations can be weighted and summed according to each weight, and the order to be allocated can be assigned to the user corresponding to the minimum weighted sum.

[0157] It should be noted that all actions involving the acquisition of signals, information, or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where the application is located, and with the authorization granted by the owner of the relevant device.

[0158] The supply and demand level determination method provided in this manual can be applied to the order allocation process where orders are assigned to delivery personnel, such as in scenarios where delivery personnel handle on-demand deliveries of express packages and food. When applied to the on-demand delivery field, the supply and demand level determination method in this manual can be used to determine the supply and demand level corresponding to the current supply and demand status, and a reasonable business scheduling strategy can be determined based on the supply and demand level, thereby improving the robustness of the scheduling decision-making system.

[0159] The above describes one or more embodiments of the supply and demand level determination method provided in this specification. Based on the same idea, this specification also provides a corresponding supply and demand level determination device, such as... Figure 5 As shown.

[0160] Figure 5 A schematic diagram of a supply and demand level determination device provided in this specification includes:

[0161] The classification module 500 is used to determine the feature values ​​of each feature dimension used to represent the supply and demand state at the current moment, as well as the classification corresponding to each feature dimension;

[0162] The membership determination module 502 is used to determine the membership degree of the feature value of each feature dimension to each category according to the category corresponding to the feature dimension, and obtain the membership degree vector corresponding to the feature dimension.

[0163] The classification combination determination module 504 is used to arrange and combine the classifications of each feature dimension to obtain each classification combination. Based on each classification combination and the membership vector corresponding to each feature dimension, the module determines the membership degree and uncertainty measure of the supply and demand state at the current moment to each classification combination.

[0164] The supply and demand level determination module 506 is used to determine the supply and demand level corresponding to the current supply and demand state based on the preset correspondence between each category combination and the supply and demand level, the membership degree of the current supply and demand state to each category combination, and the uncertainty measure, so as to determine the business scheduling strategy based on the obtained supply and demand level.

[0165] Optionally, the classification module 500, for each feature dimension, when the feature dimension is a continuous variable, obtains the historical feature values ​​of the feature dimension in a preset historical number of days, sorts the historical feature values, determines the historical feature values ​​corresponding to each percentile index based on preset percentiles and the sorting, and uses these as the classification thresholds corresponding to the feature dimension. Based on the classification thresholds corresponding to the feature dimension, it determines the classification threshold intervals corresponding to the feature dimension, wherein at least some of the classification threshold intervals overlap. Based on the classification threshold intervals corresponding to the feature dimension, it determines the classification corresponding to the feature dimension. When the feature dimension is a discrete variable, it determines the classification corresponding to the feature dimension based on the preset classification thresholds.

[0166] Optionally, the membership determination module 502 determines the membership degree of the feature value of the feature dimension to each category based on the category corresponding to the feature dimension, as a coarse classification result. It then determines whether the feature dimension is a preset feature dimension that needs to be further classified. If so, it reclassifies the feature dimension based on at least some of the fine classification thresholds corresponding to the feature dimension. Based on the categories determined after the fine classification of the feature dimension, it determines the membership degree of the feature value of the feature dimension to each category, and obtains the membership degree vector corresponding to the feature dimension. If not, it determines the membership degree vector corresponding to the feature dimension based on the coarse classification result.

[0167] Optionally, the classification combination determination module 504 determines the classification corresponding to the membership degree that meets the preset conditions for each feature dimension, and uses it as the effective classification for that feature dimension. The effective classifications of each feature dimension are then arranged and combined to obtain each classification combination.

[0168] Optionally, the classification combination determination module 504, for each classification combination, determines the membership degree of the feature value of each feature dimension to the corresponding category in the classification combination based on the membership degree vector corresponding to each feature dimension; determines the membership degree of the supply and demand state at the current moment to the classification combination based on the product of the membership degrees of the feature values ​​of each feature dimension to the corresponding categories in the classification combination; and determines the uncertainty measure of the supply and demand state at the current moment belonging to the classification combination based on the ratio of the first cumulative sum of the differences between the membership degree of the supply and demand state at the current moment to other classification combinations and the membership degree of the supply and demand state at the current moment to the classification combination, to the second cumulative sum of the membership degrees of the supply and demand state at the current moment to all classification combinations.

[0169] Optionally, the supply and demand level determination module 506 determines the category combination corresponding to the product of the membership degree and uncertainty measure of the supply and demand state at the current moment to each category combination, and determines the supply and demand level corresponding to the category combination as the supply and demand level corresponding to the current moment's supply and demand state based on the category combination and the preset correspondence between each category combination and the supply and demand level.

[0170] Optionally, the supply and demand level determination module 506 determines the delivery path for each user to deliver the orders to be assigned based on the acquired movement information and order information of each user, as well as the order information to be assigned. For each user, based on the user's delivery path, it determines the probability distribution of the completion time of each order, and samples the completion time of each order several times to determine the order timeout situation corresponding to each sample. Based on the order timeout situation corresponding to each sample and a preset timeout evaluation function, it determines the expected timeout evaluation and conditional risk timeout evaluation for the user to complete the delivery. Based on the obtained supply and demand level, the relationship between the preset supply and demand level and the first weight corresponding to the expected timeout evaluation, and the relationship between the supply and demand level and the second weight corresponding to the conditional risk timeout evaluation, it determines the first weight and the second weight. Based on the first weight and the second weight, it performs a weighted sum of the expected timeout evaluation and the conditional risk timeout evaluation corresponding to each user, and assigns the orders to be assigned to the user whose weighted sum is minimized.

[0171] This specification also provides a computer-readable storage medium storing a computer program that can be used to execute the above-described... Figure 1 The provided method for determining supply and demand levels.

[0172] This instruction manual also provides Figure 6 The diagram shows the structure of the electronic device. Figure 6At the hardware level, the electronic device includes a processor, internal bus, network interface, memory, and non-volatile memory, and may also include other hardware required for the business operations. The processor reads the corresponding computer program from the non-volatile memory into memory and then runs it to achieve the above-mentioned functions. Figure 1 The method for determining the supply and demand levels.

[0173] Of course, in addition to the software implementation method, this specification does not exclude other implementation methods, such as the combination of hardware and software XOR logic devices, etc. In other words, the execution subject of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0174] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0175] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0176] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0177] For ease of description, the above devices are described in terms of function, divided into various units. Of course, in implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware components.

[0178] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0179] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0180] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0181] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0182] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0183] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0184] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0185] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0186] Those skilled in the art will understand that the embodiments of this specification can be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this specification may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0187] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0188] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0189] The above description is merely an embodiment of this specification and is not intended to limit this specification. Various modifications and variations can be made to this specification by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this specification should be included within the scope of the claims of this specification.

Claims

1. A method for determining supply and demand levels, characterized in that, include: Determine the feature values ​​of each feature dimension used to characterize the supply and demand state at the current moment, and the corresponding classification for each feature dimension; For each feature dimension, based on the category corresponding to that feature dimension, determine the membership degree of the feature value of that feature dimension to each category, and obtain the membership degree vector corresponding to that feature dimension. The classifications of each feature dimension are permuted and combined to obtain various classification combinations. For each classification combination, based on the membership vector corresponding to each feature dimension, the membership degree of each feature dimension's feature value to the corresponding category in that classification combination is determined. Based on the product of the membership degrees of each feature dimension's feature value to the corresponding category in that classification combination, the membership degree of the current supply and demand state to that classification combination is determined. Based on the ratio of the first cumulative sum of the differences between the membership degrees of the current supply and demand state to other classification combinations and the membership degree of the current supply and demand state to that classification combination, and the second cumulative sum of the membership degrees of the current supply and demand state to all classification combinations, the uncertainty measure of the current supply and demand state to that classification combination is determined. Based on the product of the membership degree and uncertainty measure of the supply and demand status at the current moment to each category combination, determine the category combination corresponding to the product when the product reaches its maximum; based on the category combination and the preset correspondence between each category combination and the supply and demand level, determine the supply and demand level corresponding to the category combination as the supply and demand level corresponding to the supply and demand status at the current moment, so as to determine the business scheduling strategy based on the obtained supply and demand level.

2. The method as described in claim 1, characterized in that, Determine the classification corresponding to each feature dimension, specifically including: For each feature dimension, when the feature dimension is a continuous variable, obtain the historical feature value of the feature dimension in the preset historical number of days, and sort the historical feature values. Based on the preset percentiles and the sorting, the historical feature values ​​corresponding to each percentile index are determined and used as the classification threshold corresponding to that feature dimension. Based on the classification threshold corresponding to the feature dimension, determine each classification threshold interval corresponding to the feature dimension, wherein at least some of the classification threshold intervals overlap. The category corresponding to the feature dimension is determined based on the classification threshold ranges corresponding to each feature dimension. When the feature dimension is a discrete variable, the category corresponding to the feature dimension is determined according to the preset classification threshold.

3. The method as described in claim 2, characterized in that, Based on the classification corresponding to this feature dimension, determine the membership degree of each feature value belonging to each classification, and obtain the membership vector corresponding to this feature dimension, specifically including: Based on the classification corresponding to this feature dimension, determine the membership degree of the feature value of this feature dimension to each classification, and use it as the coarse classification result; Determine whether the feature dimension is a preset feature dimension that needs to be further classified; If so, then based on at least some of the sub-classification thresholds corresponding to the feature dimension, the feature dimension is reclassified, and based on the classification determined after the sub-classification of the feature dimension, the membership degree of the feature value of the feature dimension to each classification is determined, and the membership degree vector corresponding to the feature dimension is obtained. If not, then determine the membership vector corresponding to that feature dimension based on the coarse classification result.

4. The method as described in claim 3, characterized in that, The classifications of each feature dimension are permuted and combined to obtain various classification combinations, specifically including: For each feature dimension, the membership vector corresponding to the membership degree is determined, and the classification corresponding to the membership degree that meets the preset conditions is used as the effective classification of that feature dimension. The effective classifications of each feature dimension are arranged and combined to obtain various classification combinations.

5. The method as described in claim 1, characterized in that, The business scheduling strategy is determined based on the obtained supply and demand levels, characterized in that it specifically includes: Based on the acquired movement information and order information of each user, as well as the order information to be assigned, the delivery route for each user to deliver the order to be assigned is determined; For each user, based on the user's delivery route, determine the probability distribution of the completion time of each order, and sample the completion time of each order to determine the order timeout situation corresponding to each sample; Based on the order timeout status corresponding to each sample and the preset timeout evaluation function, determine the expected timeout evaluation and conditional risk timeout evaluation for the user to complete delivery; Based on the obtained supply and demand levels, the relationship between the preset supply and demand levels and the first weight corresponding to the expected timeout evaluation, and the relationship between the supply and demand levels and the second weight corresponding to the conditional risk timeout evaluation, the first weight and the second weight are determined. The expected timeout evaluation and the conditional risk timeout evaluation corresponding to each user are weighted and summed according to the first weight and the second weight, and the orders to be allocated are assigned to the user corresponding to the one whose weighted sum is minimized.

6. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of claims 1 to 5.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method described in any one of claims 1 to 5.

Images