Parking business marketing strategy optimization method and system
By acquiring real-time user navigation destinations and parking lot occupancy rates, and using distance decay coefficients and deep reinforcement learning algorithms to generate dynamic discount strategies, this solves the problem of inaccurate and timely marketing reach in existing technologies, thereby improving parking resource matching efficiency and marketing conversion rates.
Patent Information
- Application Number
- CN202511736827.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-25
- Publication Date
- 2026-03-10
AI Technical Summary
Existing technologies cannot accurately and timely reach users with marketing based on their real-time travel history and the dynamic status of parking lots, resulting in low efficiency in matching parking resources and poor marketing conversion rates.
By acquiring real-time information on user navigation destinations and parking lot occupancy rates, and using distance decay coefficients and deep reinforcement learning algorithms to dynamically generate triple time-sensitive parameters, dynamic discount strategies corresponding to the current travel scenario are generated, and countdown pushes are executed when the user approaches the destination.
It improved the accuracy and timeliness of parking business marketing strategies, enhanced user experience, and increased the efficiency of parking resource allocation and business conversion rate.
Smart Images
Figure CN121639243A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of intelligent parking, and particularly relates to a parking business marketing strategy optimization method and system. BACKGROUND
[0002] With the continuous increase of the number of cars, the demand for urban parking continues to rise, and the existing parking lots generally adopt fixed charging modes, such as hourly charging, daily charging, monthly cards or seasonal cards. Although this mode is simple to implement, it is difficult to take into account the differentiated needs of different types of users. For example, daily high-frequency commuters are more suitable for long-term discount cards, while low-frequency users prefer to pay for parking in small amounts. The fixed mode is difficult to meet the needs of both types of users at the same time, which can lead to poor user experience and limit the further improvement of the operator's revenue. In recent years, some intelligent parking platforms have introduced online parking card sales and coupon distribution mechanisms based on the Internet. For example, common fixed-period parking cards such as 30-day unlimited-use cards, monthly discount cards, or parking coupon packages such as full-price discount coupons and deduction coupons have certain effects in improving user stickiness and usage frequency. However, their pricing and discount strategies usually rely on preset rules, lack data analysis and dynamic adaptation capabilities based on individual user behavior characteristics, and cannot effectively combine real-time occupancy status of parking lots for resource allocation, which can lead to improper timing of marketing, lack of personalization of promotional content, and other issues, reducing the conversion rate of marketing activities and leaving room for improvement in terms of user satisfaction and optimization of parking lot revenue.
[0003] Therefore, in the related art, there is a technical problem that precise and timely marketing cannot be achieved based on real-time user travel and dynamic state of parking lots, resulting in low efficiency of parking resource matching and poor marketing conversion rate. SUMMARY
[0004] The present application provides a parking business marketing strategy optimization method and system, which solves the technical problem that precise and timely marketing cannot be achieved based on real-time user travel and dynamic state of parking lots in the prior art, resulting in low efficiency of parking resource matching and poor marketing conversion rate, and achieves the technical effects of improving the precision, timeliness and user experience of parking business marketing strategies, and the scheduling efficiency of parking resources and the conversion rate of business.
[0005] The application provides a parking service marketing strategy optimization method, which comprises the following steps: acquiring a current navigation destination of a user in real time based on a navigation interface, synchronously acquiring a real-time occupancy rate of a target parking lot, and constructing a trip feature vector comprising a user identifier, a current destination, a predicted arrival time, and a frequency of the same destination in the past three days; calculating a distance decay coefficient of the current trip of the user and the target parking lot by using a destination dynamic matching engine; triggering a push strategy when the distance decay coefficient meets a preset condition; calling a corresponding trip feature vector according to the push strategy, dynamically generating three-time limit parameters of a pickup time limit, a payment time limit and a failure logic according to the real-time occupancy rate and the predicted arrival time; taking the three-time limit parameters as a generation constraint to generate a dynamic discount strategy corresponding to the current trip scenario; and performing countdown push of the dynamic discount strategy when the real-time location of the user is less than a preset distance range from the current navigation destination.
[0006] In a possible implementation, the parking service marketing strategy optimization method further performs the following processing: the pickup countdown and the payment countdown of the dynamic discount strategy are configured as operation constraints on the user side; when the dynamic discount strategy is not picked up or paid by the user within the pickup time limit or the payment time limit, the dynamic discount strategy is marked as invalid according to the failure logic, and the invalid quota is automatically released into a real-time resource pool; and the recycled allocation management of the discount resources is performed according to the updated real-time resource pool for users meeting the preset condition.
[0007] In a possible implementation, the parking service marketing strategy optimization method further performs the following processing: the distance decay coefficient is calculated as follows:
[0008] ;
[0009] wherein, the distance decay coefficient is represented by, is a preset decay constant, is a distance between the target parking lot and the current location of the user.
[0010] In a possible implementation, the parking service marketing strategy optimization method further performs the following processing: if there are multiple parking lots with distance decay coefficients meeting the preset condition near the current navigation destination, an occupancy rate difference is calculated according to the real-time occupancy rates of the multiple parking lots, the target parking lot triggering the push strategy is switched to a parking lot with a lower occupancy rate when the occupancy rate difference of any two parking lots exceeds a preset threshold, and the pickup time limit is dynamically adjusted in length based on the occupancy rate difference to perform push management.
[0011] In a possible implementation, the parking service marketing strategy optimization method further performs the following processing: the pickup time limit is calculated as follows:
[0012] ;
[0013] wherein, characterizing the pickup time limit, characterizing the basic pickup time limit, is a time limit adjustment factor, is the occupancy difference, in the peak stress scene, the pickup time limit direction is the time limit shortening direction, then through calculate the pickup time limit, in the flat peak empty high scene, the pickup time limit direction is the time limit lengthening direction, through calculate the pickup time limit.
[0014] In possible implementations, the parking business marketing strategy optimization method also performs the following processing: the dynamic discount strategy is calculated according to the real-time occupancy rate according to the following formula:
[0015] ;
[0016] wherein, characterizing the dynamic pricing strategy, is the benchmark price, is the adjustment coefficient, is the real-time occupancy rate.
[0017] In possible implementations, the parking business marketing strategy optimization method also performs the following processing: the three time limit parameters are optimized through a deep reinforcement learning algorithm, the deep reinforcement learning algorithm takes the pickup success rate, the payment success rate, the discount invalid rate and the parking space vacancy cost as the joint reward function, and takes the actual user behavior sequence as the environment input, and iteratively updates the pickup time limit length, the payment time limit width and the threshold parameter through daily retraining.
[0018] The application also provides a parking service marketing strategy optimization system, comprising: a trip feature vector construction module, configured to acquire a current navigation destination of a user based on a navigation interface in real time, and synchronously acquire a real-time occupancy rate of a target parking lot, and construct a trip feature vector comprising a user identifier, a current destination, a predicted arrival time and a frequency of the same destination in the past three days; a distance decay coefficient calculation module, configured to calculate a distance decay coefficient of a current trip of the user and the target parking lot by using a destination dynamic matching engine; a push strategy triggering module, configured to trigger a push strategy when the distance decay coefficient meets a preset condition; a time limit parameter generation module, configured to call a corresponding trip feature vector according to the push strategy, and dynamically generate three time limit parameters of a pickup time limit, a payment time limit and invalidation logic according to the real-time occupancy rate and the predicted arrival time; a dynamic discount strategy generation module, configured to generate a dynamic discount strategy corresponding to a current trip scenario by taking the three time limit parameters as generation constraints; and a countdown push module, configured to perform countdown push of the dynamic discount strategy when a real-time location of the user is less than a preset distance range from the current navigation destination.
[0019] The parking service marketing strategy optimization method and system provided by the application can acquire a user navigation destination, a target parking lot occupancy rate and a user trip feature in real time, calculate a distance decay coefficient of a user trip and a parking lot by using a dynamic matching engine, trigger push when the distance decay coefficient reaches a threshold, dynamically generate three time limit parameters comprising pickup, payment and invalidation logic based on a real-time occupancy rate and a predicted arrival time, generate a dynamic discount strategy matching a current scenario, and perform dynamic strategy push with countdown when the user approaches a destination. The technical problems of being unable to accurately and timely market based on a user real-time trip and a dynamic state of a parking lot, low parking resource matching efficiency and poor marketing conversion rate in the prior art are solved, and the technical effects of improving the accuracy, timeliness and user experience of a parking service marketing strategy, and the scheduling efficiency of parking resources and the business conversion rate are achieved. BRIEF DESCRIPTION OF DRAWINGS
[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. The flowcharts are used to illustrate the operations performed by the system according to the embodiments of the present application in the present application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously as needed. Meanwhile, other operations can be added to these processes, or a step or several steps can be removed from these processes.
[0021] Figure 1 A parking service marketing strategy optimization method flowchart is provided for the embodiments of the present application.
[0022] Figure 2A parking service marketing strategy optimization system structure schematic diagram provided by the embodiment of the application.
[0023] The reference signs are explained as follows: a travel feature vector construction module 10, a distance attenuation coefficient calculation module 20, a push strategy triggering module 30, a time limit parameter generation module 40, a dynamic discount strategy generation module 50, and a countdown push module 60. DETAILED DESCRIPTION
[0024] To further illustrate the technical means and effects adopted by the application to achieve the predetermined application purposes, the following describes the specific embodiments, structures, features and effects of the application in detail in combination with the drawings and preferred embodiments.
[0025] The embodiment of the application provides a parking service marketing strategy optimization method, as shown in the method. Figure 1 The method comprises the following steps.
[0026] In step S100, the current navigation destination of a user is acquired in real time based on a navigation interface, and the real-time occupancy rate of a target parking lot is synchronously acquired, and a travel feature vector containing a user identifier, a current destination, a predicted arrival time, and a same-destination frequency in the past three days is constructed.
[0027] Preferably, the current navigation destination of the user is acquired in real time by connecting a Gaode / Baidu map API interface and ensuring that the accuracy is less than 100 meters, the real-time occupancy rate of the target parking lot is synchronously acquired, the current parking space supply and demand situation of the parking lot is reflected, then the acquired navigation destination data and the real-time occupancy rate of the parking lot are associated with the historical behavior data and the identity information of the user, and a structured travel feature vector is generated, containing multi-dimensional data such as a user identifier, a current destination, a predicted arrival time, and a same-destination frequency in the past three days. The user identifier is used to uniquely identify the user, for example, a user ID, an encrypted account, etc. The current destination is the navigation destination coordinate acquired from the navigation interface without processing. The predicted arrival time is a timestamp data derived from the navigation interface, which is usually calculated by the navigation software according to the real-time road condition and the route, and estimates the time when the user arrives at the current destination. The same-destination frequency in the past three days is a statistical value calculated by querying the historical behavior log of the user, which searches the navigation destination set by the user in the past 72 hours in the same or similar geographical range and the number of completed trips with the current destination as the key information.
[0028] In step S200, a destination dynamic matching engine is used to calculate the distance attenuation coefficient of the current travel of the user and the target parking lot.
[0029] Preferably, the destination dynamic matching engine is a component that quantitatively evaluates the spatial correlation tightness between the user and the parking lot, and calculates a quantitative index that can reflect the probability of selection decay with distance, i.e. a distance decay coefficient, based on the dynamic trip context. Specifically, according to the current position and travel direction of the user and the position of the target parking lot, the effective distance between the two is calculated, which is usually based on the actual driving distance or the estimated driving time based on the road network, wherein the distance decay coefficient is a value between 0 and 1, which is used to represent the spatial convenience probability of the user selecting the parking lot. The higher the value, the greater the attraction of the parking lot to the user in the current trip in terms of spatial distance.
[0030] Further, step S200 further comprises that the distance decay coefficient is calculated as follows:
[0031] ;
[0032] wherein, the distance decay coefficient is represented by, is a preset decay constant, is the distance between the target parking lot and the current position of the user.
[0033] Preferably, the distance decay coefficient formula is an exponential decay function, wherein, is the effective distance between the target parking lot and the current position of the user, the preset decay constant 0.05 determines the speed of decay, which is determined by historical data fitting optimization, indicating that the selection probability decreases exponentially with each unit of distance, e is the natural constant, when the distance is 0, the distance decay coefficient is the maximum value 1, and as the distance increases, the value of the distance decay coefficient rapidly decays to 0.
[0034] Step S300, when the distance decay coefficient meets the preset condition, triggering the push strategy.
[0035] Preferably, the preset condition refers to a pre-configured distance decay coefficient judgment threshold, for example, the preset threshold is 0.7, only when the distance decay coefficient is greater than or equal to the preset threshold, it means that the parking lot is close enough or convenient for the user's current trip in terms of spatial distance, thereby having the basic geographic qualification for marketing push, if the distance decay coefficient is lower than the preset threshold, the parking lot is directly filtered out.
[0036] Step S400, according to the push strategy, calling the corresponding trip feature vector, dynamically generating three time limit parameters of pickup time limit, payment time limit and invalidity logic according to the real-time occupancy rate and the estimated arrival time.
[0037] Preferably, after triggering the push strategy, the corresponding journey feature vector formed for the user is called to create, and the real-time occupancy rate and the estimated arrival time are used for this push to dynamically generate three interrelated time-sensitive parameters, including the pickup time limit, the payment time limit, and the invalidation logic. The pickup time limit is the time window in which the user has the right to pick up the coupon, and the countdown starts from the success of the push. The default pickup time limit is preset, and the larger the occupancy rate difference, the greater the shortening range of the pickup time limit. If the user is about to arrive, the pickup time limit may be set very short, forcing the user to make an immediate decision. If the user still has a long time to arrive, they may be given more time to consider. The payment time limit is the time window in which the user must complete the parking fee payment after successfully picking up the discount. The end of the payment time limit is usually set to be a buffer time after the estimated arrival time, for example, within 5 minutes after arrival, ensuring that the user can only enjoy the discount after actually using the parking lot and completing the payment, avoiding resource locking caused by users only picking up discounts but not consuming them. The invalidation logic is a condition that explicitly specifies the invalidation of the discount strategy, which may include timeout and resource release, i.e., the strategy automatically invalidates if it is not picked up within the pickup time limit or not paid within the payment time limit. If the discount strategy is invalid, the discount amount it occupies is marked as invalid and automatically released back to the real-time resource pool for use by other users, ensuring efficient recycling of limited marketing resources. This ensures that marketing resources can be managed with high precision, ensuring that the most reasonable discounts are provided to the right users at the right time, thereby maximizing conversion rates and resource utilization efficiency. The example of the three-time parameter is shown in Table 1:
[0038] Table 1: Three-time parameter data
[0039] Scenario Pickup age Payment age Expiration logic Peak ≤ 10 minutes ≤ 5 minutes before departure Over time auto release offer amount Off-peak ≤ 30 minutes ≤ 15 minutes before departure Over time marked as "historical expiration"
[0040] Further, the pickup time limit is calculated as follows:
[0041] ;
[0042] wherein, represents the pickup time limit, represents the base pickup time limit, is the time limit adjustment factor, is the occupancy rate difference, in a peak tense scenario, the pickup time limit is in the direction of time limit shortening, then the pickup time limit is calculated by in a flat peak empty high scenario, the pickup time limit is in the direction of time limit extension, then the pickup time limit is calculated by .
[0043] Preferably, the taking time limit is the effective taking time length of the current push coupon, the basic taking time limit is the standard taking time length preset under normal or average conditions, the time limit adjustment factor is a preset weight coefficient, and the time limit adjustment factor determines the influence degree of the occupancy rate difference on the final time limit. For example, if the time limit adjustment factor is 0.1, every 10% of the occupancy rate difference will have an influence of 1 unit on the taking time limit. When the user is guided from a high-occupancy parking lot to a low-occupancy parking lot, the greater the occupancy rate difference, the more scarce the target idle parking lot resource is, and the higher the value of the diversion is. By shortening the taking time limit, a signal that the opportunity is fleeting is delivered to the user, the user is prompted to take and lock the coupon as soon as possible, and the shunting and guiding are quickly completed, and the efficiency of load balancing is improved. When the overall market is idle, the main target is to encourage consumption and reduce the decision threshold of the user, and a longer consideration time is given to the user to reduce the possibility of giving up taking due to time pressure, so as to attract hesitant users, thereby improving the overall coupon taking rate and parking lot utilization rate in the period of low demand.
[0044] Further, step S400 further includes that if there are multiple parking lots with a distance decay coefficient meeting a preset condition near the current navigation end point, an occupancy rate difference is calculated according to real-time occupancy rates of the multiple parking lots, when the occupancy rate difference of any two parking lots exceeds a preset threshold, the target parking lot of the push strategy is switched to a parking lot with a lower occupancy rate, and the length of the taking time limit is dynamically adjusted based on the occupancy rate difference, and the push management is performed.
[0045] Preferably, if there are multiple parking lots near the current navigation destination with distance attenuation coefficients meeting the preset conditions, the real-time occupancy rates of the multiple parking lots are calculated to obtain the occupancy rate difference between any two of the multiple parking lots. For example, during the morning peak period on weekdays, a user A navigates to a core business district office building, a parking lot X is 300 meters away from the current position of the user A, and the real-time occupancy rate is 90%. A parking lot Y is 800 meters away from the current position of the user A, and the real-time occupancy rate is 60%. Then, a preset occupancy rate difference threshold is set, for example, 20%. If the occupancy rate difference between any two parking lots exceeds the preset occupancy rate difference threshold, the target parking lot of the push strategy is switched to the parking lot with a lower occupancy rate, so as to achieve active flow guidance. By guiding the user to a more idle parking lot, load balancing in the region is achieved, the pressure of the peak parking lot is relieved, and the income of the low utilization rate parking lot is improved. Finally, the length of the pickup time limit is dynamically adjusted according to the occupancy rate difference, so as to finely control the marketing strategy. The length of the pickup time limit refers to the length of time for the user to decide to take the coupon. The greater the occupancy rate difference, the greater the difference between the busy and idle of the parking lots, and the greater the adjustment range. For example, the target is to guide the user to leave the parking lot with a high occupancy rate and go to the parking lot with a relatively low occupancy rate. The pickup time limit is shortened to ensure that the user makes a decision as soon as possible, thereby accelerating the conversion. If it is during the off-peak period, the overall vacancy rate is relatively high. In order to encourage users to use a certain specific parking lot, the pickup time limit is extended, and the decision-making pressure is reduced. Thus, the service capability and resource utilization rate of the entire parking area are improved, and the effectiveness of the marketing strategy is maintained through dynamic parameters.
[0046] Further, step S400 further includes that the triple time limit parameters are optimized by a deep reinforcement learning algorithm. The deep reinforcement learning algorithm takes the pickup success rate, the payment success rate, the discount invalidation rate, and the parking space vacancy cost as a joint reward function, and takes the actual user behavior sequence as an environment input. The pickup time limit length, the payment time limit width, and the threshold parameter are updated by daily retraining iteration.
[0047] Preferably, the triple time limit parameters are optimized by a deep reinforcement learning algorithm, that is, by continuously interacting with real-time reaction data of users including clicking, taking, paying, ignoring, abandoning, etc. to the promotion strategy, adjusting and updating the threshold values of key parameters such as the length of the taking time limit, the width of the payment time limit, and the distance decay coefficient. Specifically, the joint reward function is determined by weighted sum of the success rate of taking, the success rate of payment, the invalidation rate of the promotion, and the cost of parking space vacancy, wherein the weights are configured according to the importance of business indicators and historical data. The success rate of taking is used to encourage the development of a time limit strategy that can attract users to click and take, the success rate of payment is used to reward the strategy that can ultimately promote users to complete payment, the invalidation rate of the promotion refers to the punishment of the strategy that sets unreasonable and leads to the invalidation of the promotion, such as too short time limit leading to the user's failure to operate in time; the cost of parking space vacancy refers to the punishment of the strategy that fails to effectively guide vehicles to lead to parking space vacancy; and the global optimal solution is determined among multiple targets such as improving user conversion, reducing resource waste, and reducing vacancy cost.
[0048] Preferably, the actual user behavior sequence is taken as the environment input, wherein the actual user behavior sequence is collected by running with the current parameter configuration every day, collecting a large number of user complete behavior logs from receiving the push to taking to payment or abandonment. Specifically, the deep reinforcement learning algorithm analyzes the user behavior data, evaluates the joint reward score obtained under the current parameters, tries to make a small, random adjustment to the parameters to explore the possibility of obtaining a higher reward, and then updates the internal neural network model based on the analysis results to establish an accurate mapping relationship between the scene state and the action parameter adjustment and the reward, and then generates an updated and more optimal parameter configuration, including the length of the taking time limit, the width of the payment time limit, and the threshold parameter, and starts a new reinforcement learning cycle in the next day, thereby ensuring adaptation to market changes such as holiday traffic mode changes, new parking lot opening, etc., continuously improving the efficiency of the overall marketing strategy and resource scheduling, and ultimately maximizing the revenue.
[0049] Step S500, generating a dynamic promotion strategy corresponding to the current trip scene by taking the triple time limit parameters as generation constraints.
[0050] Step S500 further includes that the dynamic promotion strategy is calculated according to the real-time occupancy rate according to the following formula:
[0051] ;
[0052] Wherein, represents the dynamic pricing strategy, is the benchmark price, is the adjustment coefficient, is the real-time occupancy rate.
[0053] Preferably, the triple aging parameter is taken as a generation constraint to set the effective boundary rule for the finally generated preferential policy, that is, based on the current journey scenario, the optimal preferential scheme is calculated and combined according to the real-time occupancy rate within the aging constraint, wherein the benchmark price is the historical or standard price of the parking lot, and the premium only occurs when the real-time occupancy rate exceeds 70%, that is, when the real-time occupancy rate is less than 70%, the price still maintains the benchmark price; the adjustment coefficient determines the strength of the premium, for example, if the adjustment coefficient is 0.5, when the occupancy rate reaches 90%, that is, exceeds the critical point by 20%, the price will be floated by 10%, and then the final dynamic price is calculated. Then the dynamic pricing strategy is correspondingly bound with the collection aging, payment aging and invalidation logic, and the dynamic preferential policy corresponding to the current journey scenario is output, which contains scenario-based pricing and scenario-based aging, ensuring that each marketing strategy pushed out not only has reasonable price, but also has perfect match with the real-time decision window of the user in time rhythm, so as to maximize the marketing effect.
[0054] Step S600, when the real-time position of the user is less than the preset distance range from the current navigation endpoint, the countdown push of the dynamic preferential policy is performed.
[0055] Preferably, the preset distance range is a preset accurate space threshold, for example, 800 meters, and then the final decision area centered on the destination is defined, when the physical distance between the real-time position of the user and the current navigation endpoint is less than the preset distance range, the countdown push of the dynamic preferential policy is performed, that is, the operable dynamic preferential policy message is sent to the mobile device of the user, and a visual and decreasing timer is attached, which intuitively displays the remaining time of the collection aging of the preferential policy, for example, the countdown pop-up window displays: "Collect within 8 minutes, valid before leaving, expired and invalid!"
[0056] Further, step S600 further includes step S610 of configuring the collection countdown and payment countdown of the dynamic preferential policy as the operation constraint of the user side; step S620, when the dynamic preferential policy is not collected or paid by the user within the collection aging or payment aging, the dynamic preferential policy is marked as invalid according to the invalidation logic, and the invalid quota is automatically released into the real-time resource pool; step S630, according to the updated real-time resource pool, the recirculation redistribution management of preferential resources is performed for the user satisfying the preset condition.
[0057] Preferably, the take-up countdown and the payment countdown of the dynamic discount strategy are configured, wherein the take-up countdown is the remaining time from the push for the user to have the right to take up the discount, and the payment countdown is the remaining time from the successful taking up of the user to complete the order payment, and the take-up countdown and the payment countdown are clearly displayed on the user interface to constitute the operation constraints that the user must comply with. The state of each discount strategy is monitored in real time, and if the dynamic discount strategy is not taken up or paid by the user within the taking-up time limit or the payment time limit, that is, the user ignores the push or hesitates to cause the take-up countdown to be zero, or the user takes up the discount but does not complete the payment within the specified time, then according to the invalidation logic, the dynamic discount strategy is marked as invalid in the database, and the dynamic discount strategy cannot be used by the original user again, and the corresponding invalid quota of the dynamic discount strategy is recalculated and automatically released back to the real-time resource pool. Finally, according to the updated real-time resource pool, a new user who meets the preset condition is recaptured, that is, when a new user performs matching, available discount resources are obtained from the updated real-time resource pool, and a new discount strategy is dynamically generated and pushed for the new user who meets the preset condition. Thus, the marketing budget or parking space resources are avoided from being deadlocked or wasted due to user abandonment, and at the same time, the strategy of guiding the user to the low-occupancy parking lot is perfectly matched, ensuring that each unit of marketing resource has the maximum possibility of being finally consumed, thereby improving the resource utilization rate and the comprehensive conversion rate and income of parking resources.
[0058] In the foregoing, with reference to Figure 1 A parking business marketing strategy optimization method according to an embodiment of the application is described in detail. Next, with reference to Figure 2 A parking business marketing strategy optimization system according to an embodiment of the application is described.
[0059] The parking business marketing strategy optimization system according to the embodiment of the application is used to solve the technical problem in the prior art that precise and timely marketing touch cannot be made based on the real-time journey of the user and the dynamic state of the parking lot, resulting in low parking resource matching efficiency and poor marketing conversion rate, and achieves the technical effects of improving the precision, timeliness and user experience of the parking business marketing strategy, and the scheduling efficiency and business conversion rate of the parking resources. As shown in Figure 2 A parking business marketing strategy optimization system includes a journey feature vector construction module 10, a distance decay coefficient calculation module 20, a push strategy triggering module 30, a time limit parameter generation module 40, a dynamic discount strategy generation module 50, and a countdown push module 60.
[0060] The trip feature vector construction module 10 is configured to obtain a current navigation destination of a user in real time based on a navigation interface, and synchronously obtain a real-time occupancy rate of a target parking lot, and construct a trip feature vector containing a user identifier, a current destination, a predicted arrival time, and a frequency of the same destination in the past three days; the distance decay coefficient calculation module 20 is configured to calculate a distance decay coefficient of a current trip of the user and the target parking lot by using a destination dynamic matching engine; the push strategy triggering module 30 is configured to trigger a push strategy when the distance decay coefficient meets a preset condition; the time limit parameter generation module 40 is configured to call a corresponding trip feature vector according to the push strategy, and dynamically generate three time limit parameters of a pickup time limit, a payment time limit and invalidation logic according to the real-time occupancy rate and the predicted arrival time; the dynamic discount strategy generation module 50 is configured to generate a dynamic discount strategy corresponding to a current trip scenario by taking the three time limit parameters as generation constraints; and the countdown push module 60 is configured to perform countdown push of the dynamic discount strategy when a real-time location of the user is less than a preset distance range from the current navigation destination.
[0061] In the following, the specific configuration of the countdown push module 60 will be described in detail. The countdown push module 60 further comprises: the pickup countdown and the payment countdown of the dynamic discount strategy are configured as operation constraints on the user side; when the dynamic discount strategy is not picked up or paid by the user within the pickup time limit or the payment time limit, the dynamic discount strategy is marked as invalid according to the invalidation logic, and the invalid quota is automatically released into a real-time resource pool; and the recycled allocation management of the discount resource is performed according to the updated real-time resource pool for re-capturing of the user meeting the preset condition.
[0062] In the following, the specific configuration of the distance decay coefficient calculation module 20 will be described in detail. The distance decay coefficient calculation module 20 further comprises: the distance decay coefficient is calculated as follows:
[0063] ;
[0064] wherein, the distance decay coefficient is represented by, is a preset decay constant, is a distance between the target parking lot and a current location of the user.
[0065] In the following, the specific configuration of the time limit parameter generation module 40 will be described in detail. The time limit parameter generation module 40 further comprises: if there are multiple parking lots with distance decay coefficients meeting the preset condition near the current navigation destination, an occupancy rate difference is calculated according to real-time occupancy rates of the multiple parking lots, the target parking lot triggering the push strategy is switched to a parking lot with a lower occupancy rate when the occupancy rate difference of any two parking lots exceeds a preset threshold, and the length of the pickup time limit is dynamically adjusted based on the occupancy rate difference to perform push management.
[0066] Below, the specific configuration of the time limit parameter generation module 40 will be described in detail. The time limit parameter generation module 40 further comprises: the pickup time limit is calculated as follows:
[0067]
[0068] wherein, characterizes the pickup time limit, characterizes the basic pickup time limit, is the time limit adjustment factor, is the occupancy rate difference, in the peak stress scene, the pickup time limit direction is the time limit shortening direction, then the pickup time limit is calculated by , in the flat peak empty high scene, the pickup time limit direction is the time limit lengthening direction, the pickup time limit is calculated by .
[0069] Below, the specific configuration of the dynamic discount strategy generation module 50 will be described in detail. The dynamic discount strategy generation module 50 further comprises: the dynamic discount strategy is calculated according to the real-time occupancy rate as follows:
[0070]
[0071] wherein, characterizes the dynamic pricing strategy, is the benchmark price, is the adjustment coefficient, is the real-time occupancy rate.
[0072] Below, the specific configuration of the time limit parameter generation module 40 will be described in detail. The time limit parameter generation module 40 further comprises: the triple time limit parameters are optimized by a deep reinforcement learning algorithm, the deep reinforcement learning algorithm takes the pickup success rate, the payment success rate, the discount invalid rate and the parking space vacancy cost as a joint reward function, and takes the actual user behavior sequence as an environment input, and iteratively updates the pickup time limit length, the payment time limit width and the threshold parameter through daily retraining.
[0073] The parking business marketing strategy optimization system provided by the embodiment of the application can execute the parking business marketing strategy optimization method provided by the embodiment of the application, has the function modules and beneficial effects corresponding to the execution method.
[0074] The above merely describes the preferred embodiments of the present application, and is not intended to limit the present application in any form. Although the present application has been disclosed with the preferred embodiments as above, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to the above disclosed technical content to obtain equivalent embodiments with equivalent changes, as long as the changes or modifications do not deviate from the technical solution of the present application. Any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application still belongs to the scope of the technical solution of the present application.
Claims
1. A parking business marketing strategy optimization method characterized by, The method comprises: real-time acquisition of a current navigation end point of a user based on a navigation interface, and synchronous acquisition of real-time occupancy of a target parking lot, construction of a trip feature vector containing a user identifier, a current destination, an estimated arrival time, and a same-destination frequency in the past three days; calculation of a distance decay coefficient of a current trip of the user and the target parking lot by using a destination dynamic matching engine; triggering of a push strategy when the distance decay coefficient meets a preset condition; according to the push strategy, calling a corresponding trip feature vector, and dynamically generating three-time parameters of a pickup time limit, a payment time limit, and invalidation logic according to the real-time occupancy and the estimated arrival time; generating a dynamic discount strategy corresponding to a current trip scenario by taking the three-time parameters as generation constraints; when a real-time location of the user is less than a preset distance range from the current navigation end point, performing countdown push of the dynamic discount strategy.
2. The parking business marketing strategy optimization method of claim 1, wherein, The countdown push of the dynamic discount strategy comprises: configuring a pickup countdown and a payment countdown of the dynamic discount strategy as operation constraints on the user side; when the dynamic discount strategy is not picked up or paid by the user within the pickup time limit or the payment time limit, marking the dynamic discount strategy as invalid according to the invalidation logic, and automatically releasing an invalid quota to a real-time resource pool; according to the updated real-time resource pool, re-capturing users meeting a preset condition, and performing cyclic redistribution management of discount resources.
3. The parking business marketing strategy optimization method of claim 1, wherein, The distance decay coefficient is calculated as follows: ； wherein, characterizing a distance attenuation coefficient, is a preset attenuation constant, is a distance between the target parking lot and the current location of the user.
4. The parking business marketing strategy optimization method of claim 1, wherein, if there are multiple parking lots with distance decay coefficients meeting a preset condition near the current navigation end point, calculating an occupancy difference value according to real-time occupancies of the multiple parking lots, and when an occupancy difference value of any two parking lots exceeds a preset threshold, switching a target parking lot triggering the push strategy to a parking lot with a lower occupancy, and dynamically adjusting a length of a pickup time limit based on the occupancy difference value, and performing push management.
5. The parking business marketing strategy optimization method of claim 4, wherein, The pickup time limit is calculated as follows: ； wherein, characterizing the pickup time limit, characterizing the base pickup time limit, is a time limit adjustment factor, is the occupancy difference, in a peak stress scenario, the pickup time limit direction is the time limit shortening direction, then through calculating the pickup time limit, in a flat peak empty high scenario, the pickup time limit direction is the time limit lengthening direction, through calculating the pickup time limit.
6. The parking business marketing strategy optimization method of claim 1, wherein, The dynamic discount strategy is calculated according to real-time occupancy according to the following formula: ； wherein, characterizing the dynamic pricing policy, is a reference price, is a regulation coefficient, is a real-time occupancy rate.
7. The parking business marketing strategy optimization method of claim 1, wherein, The three-time parameters are optimized by a deep reinforcement learning algorithm, the deep reinforcement learning algorithm takes a pickup success rate, a payment success rate, a discount invalidation rate, and a parking space vacancy cost as a joint reward function, and takes an actual user behavior sequence as an environment input, and iteratively updates a pickup time limit length, a payment time limit width, and threshold parameters through daily retraining.
8. A parking business marketing strategy optimization system characterized by, The system is used to implement the parking business marketing strategy optimization method in any one of claims 1 to 7, and the system comprises: a trip feature vector construction module configured to acquire a current navigation end point of a user based on a navigation interface, and synchronously acquire real-time occupancy of a target parking lot, and construct a trip feature vector containing a user identifier, a current destination, an estimated arrival time, and a same-destination frequency in the past three days; a distance decay coefficient calculation module configured to calculate a distance decay coefficient of a current trip of the user and the target parking lot by using a destination dynamic matching engine; a push strategy triggering module configured to trigger a push strategy when the distance decay coefficient meets a preset condition; The aging parameter generation module is configured to call a corresponding trip feature vector according to the push strategy, and dynamically generate three aging parameters of a pickup aging, a payment aging, and an invalidation logic according to the real-time occupancy and the predicted arrival time. The dynamic discount strategy generation module is configured to generate a dynamic discount strategy corresponding to a current trip scenario by taking the three aging parameters as generation constraints. The countdown push module is configured to perform countdown push of the dynamic discount strategy when a real-time location of the user is less than a preset distance range from the current navigation endpoint.