Recycling management and control method and system for cyclically leased carrier and electronic equipment

Abstract

The invention provides a recycling management and control method and system for cyclically leasing carriers and electronic equipment, and relates to the technical field of carrier leasing, and the method comprises the steps: inputting the obtained number of required carriers in a plurality of historical periods into a prediction model, and obtaining the number of required carriers in the next period; inputting the acquired overdue use times, the execution penalty average value and the order quantity of the user into a penalty perception model to obtain the penalty perception degree of the user; and inputting the predicted number of required carriers and the penalty perceptibility of the user into a target function, determining operation decision parameters by taking maximization of the target function as a target, and scheduling the carriers based on the operation decision parameters. By integrating the required number of the carriers in the next period, the sensitive degree of the user to punishment and the maximization of the objective function, the formulated operation decision parameters are more reasonable, the number of the carriers actively scheduled by the platform better meets the requirements of the next period, the scheduling of the carriers is optimized, and the recycling efficiency of the carriers is improved.

Description

Technical Field

[0001] The present invention relates to the field of vehicle leasing technology, and in particular to a recovery control method, system and electronic equipment for a circulating leasing vehicle. Background Art

[0002] In supply chain management, by recycling packaging materials, companies can significantly reduce waste in operations and their negative impact on the environment. Reusable carriers, as a reusable packaging, are widely used in the logistics process from primary packaging to tertiary packaging, covering a variety of types such as pallets, boxes, containers and carts. These carriers reduce space waste in warehouses and transportation in a business-to-business environment, achieving cost savings in long-term operations. In recent years, some third-party operators have established equipment rental platforms to provide customized services including rental, cleaning, warehousing and transportation coordination. This sharing economy model promotes resource sharing and collaborative use between supply chains and different industries.

[0003] Operators of the above platforms usually need to make large-scale investments to obtain vehicle resources. The platform signs a dynamic leasing agreement with the user, but due to unstable production demand or insufficient human resources, the user often lacks the motivation to return the vehicle in time at the end of the agreement, resulting in frequent vehicle shortages. This requires the platform to reserve enough vehicles in advance, forcing the platform to invest a large amount of money to compensate for the cost of default, and even temporarily lease equipment from other platforms, which hinders the efficient reuse of resources and the improvement of economic benefits.

[0004] Some existing technologies introduce digital infrastructure such as wireless radio frequency identification to enhance real-time monitoring of equipment assets and alleviate inventory shortages by optimizing reverse logistics and inventory handling. However, these methods are not efficient in scheduling vehicles, resulting in low efficiency in the recycling of vehicles. Summary of the invention

[0005] The problem to be solved by the present invention is that the existing method has low efficiency in dispatching vehicles and low efficiency in recycling vehicles.

[0006] In order to solve the above problems, in a first aspect, the present invention provides a recycling management and control method for a cyclically leased vehicle, comprising:

[0007] Inputting the acquired vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data;

[0008] The acquired number of overdue usage, average execution penalty and order volume of the user are input into the penalty perception model to obtain the user's penalty perception;

[0009] The predicted number of vehicle demands and the user's penalty perception are input into the objective function, and the operation decision parameters are determined with the goal of maximizing the objective function, wherein the operation decision parameters meet the preset constraints and include the number of vehicles actively dispatched by the platform, the number of new vehicles introduced by the platform, and the penalty intensity of each user;

[0010] The vehicle is dispatched based on the operational decision parameters.

[0011] Optionally, the training method of the penalty perception model includes:

[0012] Input the acquired user's historical overdue usage times, historical execution penalty average value and historical order volume into the initial penalty perception model to obtain the user's penalty perception;

[0013] The penalty perception of the user is input into a likelihood function, and a maximum likelihood estimation method is used to determine model parameters of an initial penalty perception model, thereby obtaining the penalty perception model.

[0014] Optionally, the penalty perception model is:

[0015]

[0016] Wherein, Y is the target variable, indicating the situation where the user chooses to execute the penalty, Y=1 indicates that the user chooses to execute the penalty, and Y=0 indicates that the user chooses not to execute the penalty; X is a feature matrix, each row represents a user feature vector, and each user feature vector includes the number of overdue uses of the user, the average execution penalty value, and the order volume; P(Y=1|X) is the probability that the user chooses to execute the penalty under the feature matrix X, that is, the user's penalty perception; X i1 represents the number of overdue uses in the i-th row of the feature matrix X, X i2 represents the average execution penalty in the i-th row of the feature matrix X, X i3 represents the order quantity in the i-th row in the feature matrix X, β0 represents the model correction parameter, β1 represents the model parameter corresponding to the number of overdue uses, β2 represents the model parameter corresponding to the average execution penalty value, and β3 represents the model parameter corresponding to the order quantity.

[0017] Optionally, the likelihood function is:

[0018]

[0019] Where L(β|X,Y) represents the probability of m events occurring when the feature matrix is ​​X, the user's option to perform penalty is Y, and the parameter of the penalty perception model is β; m represents the number of events, Y represents the probability of m events occurring. iIndicates the situation in which the user chooses to execute the penalty in the i-th event, X i represents the feature matrix in the i-th event, (P(Y i =1|X i ,β) means that when the parameter of the penalty perception model is β, in the feature matrix X i The user's perception of punishment under the circumstances.

[0020] Optionally, the objective function is:

[0021]

[0022] Among them, Max Pro represents the value of the objective function, k represents the number of the platform warehouse, K represents the number of platform warehouses, and C k represents the rental of a single vehicle in the k-th platform warehouse, D k represents the number of vehicles required by the kth platform warehouse in the next cycle, n represents the number of user warehouses, N represents the number of user warehouses, and nk represents the distance from the nth user warehouse to the kth platform warehouse, M represents the unit transportation cost, and Z nk represents the number of vehicles that the k-th platform warehouse actively dispatches from the n-th user warehouse, R nk represents the amount of vehicles returned from the nth user warehouse to the kth platform warehouse as a penalty incentive, CP k represents the cost of introducing a new vehicle to the k-th platform warehouse, P k represents the amount of new vehicles introduced into the platform of the k-th platform warehouse, CW k represents the storage cost of each vehicle in the k-th platform warehouse, IE k represents the inventory of platform vehicles in the kth platform warehouse, PP n Indicates the penalty for overdue vehicles in the nth customer warehouse, IE n Represents the inventory of user vehicles in the nth customer warehouse.

[0023] Optionally, the recycling control method of the cycle rental vehicle further includes: setting 6 preset constraints, and finding the optimal operation decision parameters under the agreement of the 6 constraints, wherein the preset constraints include:

[0024] The sum of the platform vehicle inventory in the previous period, the number of vehicles actively dispatched by the platform, the number of vehicles returned by penalty incentives, and the number of new vehicles introduced by the platform is the vehicle inventory, which can be expressed as:

[0025]

[0026] in, represents the inventory of platform vehicles in the kth platform warehouse in the previous period;

[0027] The vehicle inventory is greater than or equal to the vehicle demand quantity for the next cycle, and less than or equal to the maximum capacity of the platform warehouse. The expression is:

[0028]

[0029] in, represents the maximum capacity of the k-th platform warehouse;

[0030] The amount of vehicles returned by the penalty incentive is expressed as:

[0031]

[0032] The penalty for each user is less than or equal to the total rental amount of the user's order, expressed as:

[0033]

[0034] Among them, AC n Indicates the total rental amount of the order of the nth user;

[0035] The amount of vehicles returned by each user through penalty incentives is less than or equal to the user's vehicle inventory, expressed as:

[0036]

[0037] All the operational decision parameters are non-negative numbers, and the expressions are:

[0038]

[0039] Optionally, the predicted vehicle demand quantity and the user's penalty perception are input into an objective function, and after determining the operation decision parameters with the objective of maximizing the objective function, the method further includes:

[0040] Sending the user's penalty intensity to a corresponding user terminal, wherein the user terminal uploads feedback information on the penalty intensity, and the feedback information includes whether the user accepts the penalty intensity;

[0041] If the feedback information is received that the user does not accept the severity of the punishment, the operational decision parameters are re-determined.

[0042] Optionally, if the feedback information is received that the user does not accept the penalty intensity, re-determining the operation decision parameter includes:

[0043] If the feedback information is received that the user does not accept the penalty intensity, a new constraint condition is added, wherein the new constraint condition is that the penalty intensity to be re-determined next time is less than the currently determined penalty intensity;

[0044] Taking the maximization of the objective function as the goal, an operation decision parameter is determined, wherein the operation decision parameter satisfies a preset constraint condition and an added new constraint condition.

[0045] In a second aspect, the present invention further provides a recycling management and control system for a cyclically leased vehicle, comprising:

[0046] A demand prediction module, used to input the vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data;

[0047] A user perception prediction module is used to input the acquired number of overdue usage, average execution penalty and order volume of the user into the penalty perception model to obtain the user's penalty perception;

[0048] A decision optimization module, used to input the predicted number of vehicle demands and the user's penalty perception into an objective function, and determine the operation decision parameters with the goal of maximizing the objective function, wherein the operation decision parameters satisfy preset constraints, and the operation decision parameters include the number of vehicles actively dispatched by the platform, the number of vehicles returned as penalty incentives, the number of new vehicles introduced by the platform, and the penalty intensity of each user;

[0049] A vehicle scheduling module is used to schedule the vehicle based on the operation decision parameters.

[0050] In a third aspect, the present invention provides an electronic device, including a memory and a processor;

[0051] The memory is used to store computer programs;

[0052] The processor is used to implement the recycling management and control method of the recycled rental vehicle as described above when executing the computer program.

[0053] The beneficial effects of the recycling control method, system and electronic equipment of the circular rental vehicle of the present invention are:

[0054] By inputting the number of vehicle demands obtained in multiple historical cycles into the prediction model, the number of vehicle demands in the next cycle can be obtained, so that the number of vehicle demands in the future can be predicted in advance, which is convenient for the platform to check in advance whether the vehicle inventory in the current warehouse is sufficient and plan the vehicle scheduling plan in advance; the number of overdue uses, the average execution penalty and the order volume of the user are input into the penalty perception model to obtain the user's penalty perception, and the probability of the user choosing to execute the penalty can reflect the user's sensitivity to the currently formulated penalty, which can assist the platform in formulating the appropriate penalty intensity, ensure user satisfaction and maintain The order volume of the platform can also improve the rationality of the penalty formulation plan and optimize the scheduling of inventory; the predicted number of vehicle demands and the user's penalty perception are input into the objective function, and the objective function is maximized to determine the number of vehicles that the platform actively dispatches to meet the preset constraints, the number of vehicles returned by penalty incentives, the number of new vehicles introduced by the platform, and the penalty intensity of each user. Based on the operational decision parameters, the vehicles are scheduled. For example, the corresponding number of vehicles is dispatched from other platform warehouses according to the number of new vehicles introduced by the platform. The operational decision parameters can be quickly determined through the model, and the efficiency of vehicle scheduling based on the operational decision parameters can be improved. The number of vehicle demands in the next cycle, the user's sensitivity to penalties, and the goal of maximizing the objective function make the final operational decision parameters more reasonable, which can not only ensure user satisfaction with the platform, but also stimulate users to actively return vehicles on time to improve the recycling efficiency of vehicles. The number of vehicles actively dispatched by the platform is more in line with the needs of the next cycle, so that the scheduling of vehicles is optimized, which is conducive to the continuous and healthy operation of the platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A schematic diagram of a flow chart of a recycling and leasing vehicle recovery and control method provided by an embodiment of the present invention;

[0056] Figure 2 A schematic diagram of the structure of a recycling and leasing vehicle recovery and control system provided by an embodiment of the present invention;

[0057] Figure 3 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be interpreted as being limited to the embodiments described herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not intended to limit the scope of protection of the present invention.

[0059] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0060] The term "including" and its variations used in this document are open inclusions, that is, "including but not limited to"; the term "based on" means "based at least in part on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiments". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc. mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0061] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0062] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes, and are not used to limit the scope of these messages or information.

[0063] In response to the problems existing in the above-mentioned related technologies, this embodiment provides a recovery control method, system and electronic equipment for a circular rental vehicle.

[0064] like Figure 1 As shown, an embodiment of the present invention provides a recycling management and control method for a cyclic rental vehicle, including:

[0065] S1: Input the acquired vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data.

[0066] Specifically, the neural network model can be a long short-term memory network model, or a bidirectional long short-term memory network model. The above-mentioned cycle can be a decision cycle, and the length of the cycle can be 1 week, 1 month, 2 months or half a year, etc. The length of the cycle is set according to actual usage requirements. Before the preset model is put into use, it is necessary to use historical data to train the neural network model, collect the number of vehicle requirements in the platform warehouse in the past multiple cycles, and perform pre-processing operations such as cleaning, feature engineering and standardization on the collected data; then divide the data into training set, validation set and test set; select appropriate activation function and loss function, and train the model through the training set. For example, the number of vehicles required for a platform warehouse from the 1st to the 10th cycle can be obtained; the number of vehicles required for the platform warehouse in each cycle from the 1st to the 3rd cycle can be used as the input value of the neural network model, and then the neural network model outputs a predicted value, which is compared with the number of vehicles required for the platform warehouse in the 4th cycle, and the parameters of the neural network model are adjusted according to the comparison results. Similarly, the number of vehicles required for the platform warehouse in the past one or more cycles can be used as the input value of the neural network model, and then the number of vehicles required for the platform warehouse in the next cycle can be predicted, and the predicted value can be compared with the actual demand quantity, and the parameters of the neural network model can be continuously adjusted until the loss value between the predicted value and the actual value of the neural network model is within an acceptable range, and the training of the neural network model is stopped to obtain the prediction model. Finally, the validation set and test set are used for verification and testing, and the final tuning is performed. The number of vehicle requirements for multiple historical cycles is input into the prediction model, so that the number of vehicle requirements for the next cycle can be obtained, and the number of vehicle requirements for each platform warehouse in the next cycle can be predicted in advance, so that the total demand for vehicles in all platform warehouses can be known, which is convenient for the platform to schedule and allocate vehicles in advance. For example, if the demand for vehicles in the platform warehouse in the next cycle is not high or the current inventory quantity can meet the demand, the punishment for users can be relatively relaxed in the following penalty intensity formulation, and the punishment can be a penalty measure such as limiting the maximum value of subsequent vehicle rentals, limiting the term of subsequent vehicle rentals, or increasing the penalty amount for overdue. This method can prevent the platform from losing more costs, and can also ensure user satisfaction and maintain the order volume of the platform. Therefore, using the prediction model to accurately predict the number of vehicle requirements during the future decision-making period can improve the rationality of the penalty intensity formulation plan, optimize the scheduling of inventory, and improve scheduling efficiency. In addition, real-time data can be continuously collected, and the prediction model can be updated and maintained as needed.

[0067] S2: Input the acquired number of overdue usage, average execution penalty and order volume of the user into the penalty perception model to obtain the user's penalty perception.

[0068] Specifically, in the actual operation scenario of the recycling vehicle, it is necessary to comprehensively consider the user's historical behavior data, such as the number of historical overdue uses, the user's sensitivity to penalties, etc. Through data mining, the user's transaction records, historical selection behaviors, and related statistical information are collected to extract key features that affect the user's selection behavior, such as the historical penalty intensity, the user's historical order volume, and the number of historical overdue uses.

[0069] For example, let X i is a feature matrix, where each row represents a user's feature vector, then X i1 represents the number of historical overtime usage of the user on the platform in the i-th feature matrix, X i2 represents the average historical execution penalty of the user in the i-th feature matrix, X i3 Represents the historical order volume of the user in the i-th feature matrix on the platform. Let Y be the target variable, indicating whether the user chooses to execute the penalty. It is usually a binary classification problem with a value of 0 or 1. Y = 1 means that the user chooses to execute the penalty, and Y = 0 means that the user chooses not to execute the penalty. Select a suitable machine learning algorithm to build a penalty perception model. The penalty perception model is:

[0070]

[0071] Wherein, Y is the target variable, indicating the situation where the user chooses to execute the penalty, Y=1 indicates that the user chooses to execute the penalty, and Y=0 indicates that the user chooses not to execute the penalty; X is a feature matrix, each row represents a user feature vector, and each user feature vector includes the number of overdue uses of the user, the average execution penalty value, and the order volume; P(Y=1|X) is the probability that the user chooses to execute the penalty under the feature matrix X, that is, the user's penalty perception; X i1 represents the number of overdue uses in the i-th row of the feature matrix X, X i2 represents the average execution penalty in the i-th row of the feature matrix X, X i3 represents the order quantity in the i-th row in the feature matrix X, β0 represents the model correction parameter, β1 represents the model parameter corresponding to the number of overdue uses, β2 represents the model parameter corresponding to the average execution penalty value, and β3 represents the model parameter corresponding to the order quantity.

[0072] Through the constructed penalty perception model, we can know the probability of users choosing to execute penalties, which can reflect the user's sensitivity to penalties. For example, if the penalty is too high, the probability of users choosing to execute penalties will be lower, and they will choose to return the vehicle; if the penalty is too low, the probability of users choosing to execute penalties will increase, extending the use time of the vehicle. The penalty intensity can be dynamically adjusted according to the different sensitivity of different users to penalties.

[0073] S3: Input the predicted number of vehicle demands and the user's penalty perception into the objective function, and determine the operational decision parameters with the goal of maximizing the objective function, wherein the operational decision parameters satisfy preset constraints and include the number of vehicles actively dispatched by the platform, the number of vehicles returned as penalty incentives, the number of new vehicles introduced by the platform, and the penalty intensity for each user.

[0074] Specifically, after receiving the predicted number of vehicle demands and the user's penalty perception (i.e., the user's perception of penalty), the objective function and constraints are set to assist in the operation decision, and the objective function is solved by the solver to obtain the corresponding operation decision parameters. The objective function and the preset constraints are as follows.

[0075] For example, considering the optimal operating profit of the rental platform, the objective function is:

[0076]

[0077] Among them, Pro represents profit, k represents the number of platform warehouses, K represents the number of platform warehouses, and C k represents the rental of a single vehicle in the k-th platform warehouse, D k represents the number of vehicles required by the kth platform warehouse in the next cycle, n represents the number of user warehouses, N represents the number of user warehouses, and nk represents the distance from the nth user warehouse to the kth platform warehouse, M represents the unit transportation cost, and Z nk represents the number of vehicles that the k-th platform warehouse actively dispatches from the n-th user warehouse, R nk represents the amount of vehicles returned from the nth user warehouse to the kth platform warehouse as a penalty incentive, CP k represents the cost of introducing a new vehicle to the k-th platform warehouse, P k represents the amount of new vehicles introduced into the platform of the k-th platform warehouse, CW k represents the storage cost of each vehicle in the k-th platform warehouse, IE k represents the inventory of platform vehicles in the kth platform warehouse, PP n Indicates the penalty for overdue vehicles in the nth customer warehouse, IE n Represents the inventory of user vehicles in the nth customer warehouse.

[0078] Six preset constraints are set, and the optimal operation decision parameters (including the number of vehicles actively dispatched by the platform, the number of vehicles returned by penalty incentives, the number of new vehicles introduced by the platform, and the penalty intensity of each user) are found under the agreement of these six constraints. The preset constraints include:

[0079] ① The sum of the platform vehicle inventory in the previous period, the number of vehicles actively dispatched by the platform, the number of vehicles returned by penalty incentives, and the number of new vehicles introduced by the platform is the vehicle inventory, which can be expressed as:

[0080]

[0081] in, represents the inventory of platform vehicles in the kth platform warehouse in the previous period;

[0082] ② The vehicle inventory is greater than or equal to the vehicle demand in the next cycle, and less than or equal to the maximum capacity of the platform warehouse. The expression is:

[0083]

[0084] in, represents the maximum capacity of the k-th platform warehouse;

[0085] ③The amount of vehicles returned by the penalty incentive is expressed as:

[0086]

[0087] If the user chooses to execute the penalty, it means that the user chooses to continue holding the punished vehicle. Therefore, the probability that the user chooses to execute the penalty can be used as the estimated proportion of the user choosing to retain the vehicle. 1 minus the preset proportion of retained vehicles can be regarded as the return proportion of the vehicles returned through penalty incentives; therefore, the amount of vehicles returned through penalty incentives is equal to the user's vehicle inventory multiplied by the possible return proportion.

[0088] ④ The penalty for each user is less than or equal to the total rental amount of the user's order, expressed as:

[0089]

[0090] Among them, AC n Indicates the total rental amount of the order of the nth user;

[0091] ⑤ The amount of vehicles returned by each user through penalty incentives is less than or equal to the user's vehicle inventory, expressed as:

[0092]

[0093] ⑥ All the operational decision parameters are non-negative numbers, and the expression is:

[0094]

[0095] The sum of the platform's vehicle inventory in the previous cycle, the number of vehicles actively dispatched by the platform, the number of vehicles returned through penalty incentives, and the number of new vehicles introduced to the platform is greater than or equal to the number of vehicles required in the next cycle, and less than or equal to the maximum capacity of the platform warehouse; the penalty intensity for each user is less than or equal to the total rental amount of the user's order; the number of vehicles returned by each user through penalty incentives is less than or equal to the user's vehicle inventory; and all of the above-mentioned operational decision parameters are non-negative numbers.

[0096] S4: Scheduling the vehicle based on the operation decision parameters.

[0097] Specifically, according to the number of vehicles actively dispatched by the platform, the number of vehicles returned by penalty incentives, and the number of new vehicles introduced by the platform in the operational decision parameters, the vehicles in each warehouse are dispatched, or the distribution of vehicles is adjusted between warehouses, or new vehicles are introduced from other platforms, so that the platform's vehicle holdings can meet the number of vehicle requirements in the next cycle. In addition, the intensity of the penalty for each user can be adjusted to encourage users to return vehicles. When the number of vehicles returned by users increases, the number of vehicles actively dispatched by the platform and the number of new vehicles introduced by the platform can be reduced accordingly, reducing the flow of vehicles between warehouses and avoiding unnecessary dispatch. For example, the vehicle demand in the next cycle is 300, the number of vehicles returned by penalty incentives is 100, the number of vehicles actively dispatched by the platform is 150, and the remaining vehicles need to be purchased by the platform or transferred from other platforms, so the number of new vehicles introduced by the platform is 50.

[0098] In this embodiment, the vehicle demand quantity of multiple historical periods is input into the prediction model to obtain the vehicle demand quantity of the next period, so that the future vehicle demand quantity can be predicted in advance, which is convenient for the platform to check in advance whether the vehicle inventory in the current warehouse is sufficient and plan the vehicle scheduling plan in advance; the number of overdue uses, the average execution penalty and the order volume of the user are input into the penalty perception model to obtain the user's penalty perception, and the probability of the user choosing to execute the penalty can reflect the user's sensitivity to the currently formulated penalty, which can assist the platform in formulating the appropriate penalty intensity to ensure user satisfaction. degree, maintain the order volume of the platform, improve the rationality of the penalty formulation plan, and optimize the scheduling of inventory; input the predicted vehicle demand quantity and the user's penalty perception into the objective function, and take the maximization of the objective function as the goal to determine the number of vehicles actively scheduled by the platform that meet the preset constraints, the number of vehicles returned by the penalty incentive, the number of new vehicles introduced by the platform, and the penalty intensity of each user. Based on the operational decision parameters, the vehicles are scheduled. For example, according to the number of new vehicles introduced by the platform, the corresponding number of vehicles are dispatched from other platform warehouses. The model can quickly determine the operational decision parameters, and then improve the efficiency of vehicle scheduling based on the operational decision parameters. The number of vehicles required in the next cycle, the user's sensitivity to penalties, and the goal of maximizing the objective function make the final operational decision parameters more reasonable, which can not only ensure user satisfaction with the platform, but also stimulate users to actively return vehicles on time to improve the recycling efficiency of vehicles. The number of vehicles actively scheduled by the platform is more in line with the needs of the next cycle, so that the scheduling of vehicles is optimized, which is conducive to the continuous and healthy operation of the platform.

[0099] Optionally, the training method of the penalty perception model includes:

[0100] Input the acquired user's historical overdue usage times, historical execution penalty average value and historical order volume into the initial penalty perception model to obtain the user's penalty perception;

[0101] The penalty perception of the user is input into a likelihood function, and a maximum likelihood estimation method is used to determine model parameters of an initial penalty perception model, thereby obtaining the penalty perception model.

[0102] Specifically, the likelihood function is:

[0103]

[0104] Where L(β|X,Y) represents the probability of m events occurring when the feature matrix is ​​X, the user's option to perform penalty is Y, and the parameter of the penalty perception model is β; m represents the number of events, Y represents the probability of m events occurring. i Indicates the situation in which the user chooses to execute the penalty in the i-th event, X irepresents the feature matrix in the i-th event, (P(Y i =1|X i ,β) means that when the parameter of the penalty perception model is β, in the feature matrix X i The user's perception of punishment under the circumstances.

[0105] In this optional embodiment, when training the penalty-aware model, cross-validation and other techniques are used to avoid overfitting, and the model parameters are adjusted to optimize the accuracy of the results. After the model training is completed, rigorous evaluation and testing are performed to ensure its generalization ability in different scenarios. Finally, the trained model is deployed on the online platform to receive user behavior data in real time, quickly and accurately predict the user's choice of punishment, thereby helping the platform to formulate a more reasonable punishment policy and optimize the user experience.

[0106] Optionally, the predicted vehicle demand quantity and the user's penalty perception are input into an objective function, and after determining the operation decision parameters with the objective of maximizing the objective function, the method further includes:

[0107] Sending the user's penalty intensity to a corresponding user terminal, wherein the user terminal uploads feedback information on the penalty intensity, and the feedback information includes whether the user accepts the penalty intensity;

[0108] If the feedback information is received that the user does not accept the penalty intensity, the operation decision parameter is re-determined. If the feedback information is received that the user accepts the penalty intensity, the operation decision parameter is directly executed.

[0109] Specifically, if the feedback information is received that the user does not accept the penalty intensity, a new constraint condition is added, wherein the new constraint condition is that the penalty intensity re-determined next time is less than the currently determined penalty intensity; with the maximization of the objective function as the goal, the operation decision parameters are determined, wherein the operation decision parameters satisfy the preset constraints and the newly added constraints.

[0110] In this embodiment, some users cannot return the vehicle on time due to practical problems, such as the vehicle is still in use, but the use period only needs to be postponed for a few days. In this case, the user may not renew a longer rental agreement, but may choose to execute a penalty. In this case, the platform sends the penalty intensity determined by maximizing the objective function to the user. If the user feels that the penalty intensity is high, the user can provide online feedback through the user terminal. After receiving the user feedback, the platform can choose to sacrifice part of the profit and give the user a better user experience to ensure that the user will choose this platform for vehicle rental services in the future. The platform can re-determine the operation decision parameters, but the penalty intensity in the next re-determined operation decision parameters needs to be less than the currently determined penalty intensity. Due to the addition of constraints, the platform's profit is reduced under the constraints. If the platform can accept it, the re-determined penalty intensity can be sent to the user again, so as to reach an agreement with the user and enhance the user's favorability.

[0111] like Figure 2 As shown, an embodiment of the present invention provides a recycling and management system for a cyclic rental vehicle, including:

[0112] The demand prediction module 210 is used to input the vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data;

[0113] The user perception prediction module 220 is used to input the acquired number of overdue usage, average execution penalty and order volume of the user into the penalty perception model to obtain the user's penalty perception;

[0114] The decision optimization module 230 is used to input the predicted number of vehicle demands and the user's penalty perception into the objective function, and determine the operation decision parameters with the goal of maximizing the objective function, wherein the operation decision parameters meet the preset constraints, and the operation decision parameters include the number of vehicles actively dispatched by the platform, the number of vehicles returned by penalty incentives, the number of new vehicles introduced by the platform, and the penalty intensity of each user;

[0115] The vehicle scheduling module 250 is used to schedule the vehicle based on the operation decision parameters.

[0116] In this embodiment, the demand prediction module 210 inputs the vehicle demand quantity obtained in multiple historical periods into the prediction model to obtain the vehicle demand quantity in the next period, so as to be able to predict the future demand quantity for vehicles in advance, so as to facilitate the platform to check in advance whether the vehicle inventory in the current warehouse is sufficient and plan the vehicle scheduling plan in advance; the user perception prediction module 220 inputs the number of overdue uses, the average execution penalty and the order volume of the user into the penalty perception model to obtain the user's penalty perception, and the probability of the user choosing to execute the penalty can reflect the user's sensitivity to the currently formulated penalty, so as to assist the platform in formulating a suitable penalty intensity, which can ensure user satisfaction, maintain the platform's order volume, and improve the rationality of the penalty formulation plan and optimize inventory scheduling; the decision optimization module 230 inputs the predicted vehicle demand quantity and the user's penalty perception into the objective function, and takes the maximization of the objective function as the goal to determine the amount of vehicles actively scheduled by the platform that meet the preset constraints, the amount of vehicles returned by the penalty incentive, the amount of new vehicles introduced by the platform and the penalty intensity of each user; the vehicle scheduling module 250 schedules the vehicle based on the operation decision parameters. Taking into account the number of vehicles required in the next cycle, the user's sensitivity to penalties, and the goal of maximizing the objective function, the final operational decision parameters are made more reasonable. This can not only ensure user satisfaction with the platform and stimulate users to actively return vehicles on time, but also the number of vehicles actively dispatched by the platform is more in line with the needs of the next cycle, thereby optimizing the scheduling of vehicles, which is conducive to the continuous and healthy operation of the platform and improving the recycling efficiency of vehicles.

[0117] Alternatively, if Figure 2 As shown, the recycling management and control system of the recycled rental vehicle includes a customer evaluation module, and the customer evaluation module 240 is used to send the user's penalty intensity to the corresponding user terminal, wherein the user terminal uploads feedback information on the penalty intensity, and the feedback information includes whether the user accepts the penalty intensity; if the feedback information is received that the user does not accept the penalty intensity, the operation decision parameters are re-determined.

[0118] Implementing the above method on the actual rental platform, the dynamic formulation of penalties is effective in motivating users to actively return overdue vehicles. At the same time, through the user's sensitivity to penalties, it is observed that different industries have different sensitivities to penalties. For example, smaller users are more sensitive to penalties than larger users. Therefore, the sensitivity to penalties is also related to the user's industry and user scale. This also shows the necessity of the strategy of dynamically adjusting penalties in the above method in the actual operation process.

[0119] like Figure 3As shown, an electronic device 300 provided by an embodiment of the present invention includes a memory 320 and a processor 310; the memory 320 is used to store a computer program; the processor 310 is used to implement the recycling control method of the circular rental vehicle as described above when executing the computer program.

[0120] In other words, an electronic device 300 includes a memory 320 and a processor 310 coupled to the memory 320; the memory 320 is configured to store a computer program; and the processor 310 is configured to perform the following operations when executing the computer program:

[0121] Inputting the acquired vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data;

[0122] The acquired number of overdue usage, average execution penalty and order volume of the user are input into the penalty perception model to obtain the user's penalty perception;

[0123] The predicted number of vehicle demands and the user's penalty perception are input into the objective function, and the operational decision parameters are determined with the goal of maximizing the objective function, wherein the operational decision parameters satisfy preset constraints and include the number of vehicles actively dispatched by the platform, the number of vehicles returned as penalty incentives, the number of new vehicles introduced to the platform, and the penalty intensity for each user.

[0124] The technical effects of the electronic equipment in this embodiment are similar to those of the above-mentioned recycling rental vehicle recovery control method, and will not be repeated here.

[0125] An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the recycling and control method of the recycled rental vehicle as described above is implemented.

[0126] In other words, a non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor performs the following operations:

[0127] Inputting the acquired vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data;

[0128] The acquired number of overdue usage, average execution penalty and order volume of the user are input into the penalty perception model to obtain the user's penalty perception;

[0129] The predicted number of vehicle demands and the user's penalty perception are input into the objective function, and the operational decision parameters are determined with the goal of maximizing the objective function, wherein the operational decision parameters satisfy preset constraints and include the number of vehicles actively dispatched by the platform, the number of vehicles returned as penalty incentives, the number of new vehicles introduced to the platform, and the penalty intensity for each user.

[0130] The technical effects of the computer-readable storage medium in this embodiment are similar to those of the above-mentioned recycling rental vehicle recovery control method, and will not be repeated here.

[0131] Now will describe the electronic device 300 that can be used as the server or user end of the present invention, it is the example of the hardware device that can be applied to each aspect of the present invention.Electronic device 300 is intended to represent the computer equipment of various forms of digital electronics, such as, laptop computer, desktop computer, workbench, personal digital assistant, server, blade server, mainframe computer and other suitable computer.Electronic device 300 can also represent various forms of mobile device, such as, personal digital processing, cellular phone, smart phone, wearable device and other similar computing device.The parts shown in this article, their connection and relationship, and their functions are only as examples, and are not intended to limit the realization of the present invention described and / or required herein.

[0132] The electronic device 300 includes a computing unit, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) or a computer program loaded from a storage unit into a random access memory (RAM). In the RAM, various programs and data required for the operation of the device can also be stored. The computing unit, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0133] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc. In the present application, the unit described as a separate component may or may not be physically separated, and the component displayed as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment of the present invention. In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0134] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. A recycling management and control method for a cyclically leased vehicle, characterized in that: include: Inputting the acquired vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data; The acquired number of overdue usage, average execution penalty and order volume of the user are input into the penalty perception model to obtain the user's penalty perception; The predicted number of vehicle demands and the user's penalty perception are input into the objective function, and the operation decision parameters are determined with the goal of maximizing the objective function, wherein the operation decision parameters meet the preset constraints and include the number of vehicles actively dispatched by the platform, the number of new vehicles introduced by the platform, and the penalty intensity of each user; The vehicle is dispatched based on the operational decision parameters.

2. The recycling management and control method of the cycle rental vehicle according to claim 1 is characterized in that: The training method of the penalty perception model includes: Input the acquired user's historical overdue usage times, historical execution penalty average value and historical order volume into the initial penalty perception model to obtain the user's penalty perception; The penalty perception of the user is input into a likelihood function, and a maximum likelihood estimation method is used to determine model parameters of an initial penalty perception model, thereby obtaining the penalty perception model.

3. The recycling control method of the cycle rental vehicle according to claim 1 is characterized in that: The penalty perception model is: Wherein, Y is the target variable, indicating the situation where the user chooses to execute the penalty, Y=1 indicates that the user chooses to execute the penalty, and Y=0 indicates that the user chooses not to execute the penalty; X is a feature matrix, each row represents a user feature vector, and each user feature vector includes the number of overdue uses of the user, the average execution penalty value, and the order volume; P(Y=1|X) is the probability that the user chooses to execute the penalty under the feature matrix X, that is, the user's penalty perception; X i1 represents the number of overdue uses in the i-th row of the feature matrix X, X i2 represents the average execution penalty in the i-th row of the feature matrix X, X i3 represents the order quantity in the i-th row in the feature matrix X, β0 represents the model correction parameter, β1 represents the model parameter corresponding to the number of overdue uses, β2 represents the model parameter corresponding to the average execution penalty value, and β3 represents the model parameter corresponding to the order quantity.

4. The recycling control method for a cyclically leased vehicle according to claim 2, characterized in that: The likelihood function is: Where L(β|X, Y) represents the probability of m events occurring when the feature matrix is ​​X, the user's penalty option is Y, and the parameter of the penalty perception model is β; m represents the number of events, Y represents the probability of m events occurring. i Indicates the situation in which the user chooses to execute the penalty in the i-th event, X i represents the feature matrix in the i-th event, (P(Y i =1|X i , β) means that when the parameter of the penalty perception model is β, in the feature matrix X i The user's perception of punishment under the circumstances.

5. The recycling management and control method of the cycle rental vehicle according to claim 3 is characterized in that: The objective function is: Among them, Max Pro represents the value of the objective function, k represents the number of the platform warehouse, K represents the number of platform warehouses, and C k represents the rental of a single vehicle in the k-th platform warehouse, D k represents the number of vehicles required by the k-th platform warehouse in the next cycle, n represents the number of user warehouses, N represents the number of user warehouses, and nk represents the distance from the nth user warehouse to the kth platform warehouse, M represents the unit transportation cost, and Z nk represents the number of vehicles that the k-th platform warehouse actively dispatches from the n-th user warehouse, R nk represents the amount of vehicles returned from the nth user warehouse to the kth platform warehouse as a penalty incentive, CP k represents the cost of introducing a new vehicle to the k-th platform warehouse, P k represents the amount of new vehicles introduced into the platform of the k-th platform warehouse, CW k represents the storage cost of each vehicle in the k-th platform warehouse, IE k represents the inventory of platform vehicles in the kth platform warehouse, PP n Indicates the penalty for overdue vehicles in the nth customer warehouse, IE n Represents the inventory of user vehicles in the nth customer warehouse.

6. The recycling management and control method of the cycle rental vehicle according to claim 5 is characterized in that: Also includes: Six preset constraints are set, and the optimal operation decision parameters are found under the agreement of these six constraints. The preset constraints include: The sum of the platform vehicle inventory in the previous period, the number of vehicles actively dispatched by the platform, the number of vehicles returned by penalty incentives, and the number of new vehicles introduced by the platform is the vehicle inventory, which can be expressed as: in, represents the inventory of platform vehicles in the kth platform warehouse in the previous period; The vehicle inventory is greater than or equal to the vehicle demand quantity for the next cycle, and less than or equal to the maximum capacity of the platform warehouse. The expression is: in, represents the maximum capacity of the k-th platform warehouse; The amount of vehicles returned by the penalty incentive is expressed as: The penalty for each user is less than or equal to the total rental amount of the user's order, expressed as: Among them, AC n Indicates the total rental amount of the order of the nth user; The amount of vehicles returned by each user through penalty incentives is less than or equal to the user's vehicle inventory, expressed as: All the operational decision parameters are non-negative numbers, and the expressions are:

7. The recycling management and control method of the cycle rental vehicle according to claim 1 is characterized in that: The step of inputting the predicted vehicle demand quantity and the user's penalty perception into an objective function, taking maximizing the objective function as a goal, and determining the operation decision parameters further includes: Sending the user's penalty intensity to a corresponding user terminal, wherein the user terminal uploads feedback information on the penalty intensity, and the feedback information includes whether the user accepts the penalty intensity; If the feedback information is received that the user does not accept the severity of the punishment, the operational decision parameters are re-determined.

8. The recycling management and control method of the cycle rental vehicle according to claim 7 is characterized in that: If the feedback information that the user does not accept the penalty intensity is received, re-determining the operation decision parameter includes: If the feedback information is received that the user does not accept the penalty intensity, a new constraint condition is added, wherein the new constraint condition is that the penalty intensity to be re-determined next time is less than the currently determined penalty intensity; Taking the maximization of the objective function as the goal, an operation decision parameter is determined, wherein the operation decision parameter satisfies a preset constraint condition and an added new constraint condition.

9. A recycling management and control system for a cycle rental vehicle, characterized in that: include: A demand prediction module, used to input the vehicle demand quantities of multiple historical periods into a prediction model to obtain the vehicle demand quantity of the next period, wherein the prediction model is a model obtained by training a neural network model based on historical vehicle demand data; A user perception prediction module is used to input the acquired number of overdue usage, average execution penalty and order volume of the user into the penalty perception model to obtain the user's penalty perception; A decision optimization module, used to input the predicted number of vehicle demands and the user's penalty perception into an objective function, and determine the operation decision parameters with the goal of maximizing the objective function, wherein the operation decision parameters satisfy preset constraints, and the operation decision parameters include the number of vehicles actively dispatched by the platform, the number of vehicles returned as penalty incentives, the number of new vehicles introduced by the platform, and the penalty intensity of each user; A vehicle scheduling module is used to schedule the vehicle based on the operation decision parameters.

10. An electronic device, characterized in that: including memory and processor; The memory is used to store computer programs; The processor is used to implement the recycling management and control method of the recyclable rental vehicle as described in any one of claims 1 to 8 when executing the computer program.

Images