Intelligent running visit journey optimization method for terminal business personnel

Abstract

The invention discloses an intelligent running visit journey optimization method for terminal business personnel, and particularly relates to the field of visit journey optimization, and the method comprises the steps: collecting target customer state data, current execution dynamic data of a salesman and external environment change data in real time, and recognizing business abnormal dynamic data based on target journey decision data and a preset rule; calculating a comprehensive priority score of each customer based on the business abnormal dynamic data, sorting the scores from high to low, and planning an initial optimal visiting route in combination with the real-time positions and road conditions of salesmen; continuously monitoring the target travel decision data, updating the comprehensive priority score of the customer in real time when an exception is found, locking the current customer with the highest value according to the updated comprehensive priority score of the customer, and recalculating and pushing an optimized route; and pushing the optimized route to a salesman terminal, labeling the abnormal customer, and realizing a complete closed loop of data acquisition, matrix operation, feedback execution and parameter iteration.

Description

Technical Field

[0001] This invention relates to the field of visitation schedule optimization technology, and more specifically, to a method for optimizing the intelligent running visitation schedule for end-user business personnel. Background Technology

[0002] With intensifying business competition and the advancement of enterprise digitalization, the efficiency of the travel planning of field visits by front-end business personnel has a significant impact on the competitiveness of enterprises in the front-end market, as it is a key link in maintaining customer relationships and promoting business results.

[0003] Traditional visit scheduling relies on the experience of managers and the independent arrangements of sales personnel, generally resulting in problems such as arbitrary route planning, uneven customer coverage, and low time utilization efficiency. To improve this situation, intelligent management tools for field visits have emerged in the industry, enabling online assignment of visit tasks and basic route planning functions. However, in actual use, these tools still have some shortcomings: on the one hand, existing tools are mostly based on static customer information and simple geographical distances for itinerary planning, failing to integrate key elements such as the real-time status of sales personnel and dynamic changes in customer needs, leading to a disconnect between planned itineraries and actual business scenarios.

[0004] On the other hand, existing tools lack flexible adjustment mechanisms when dealing with complex changes in the field environment. When encountering emergencies such as traffic congestion or customers temporarily changing their schedules, they cannot quickly generate the best alternative solutions, making it difficult to meet the requirements of efficient and flexible operation of field work. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent running visit route optimization method for terminal business personnel, which solves the problems mentioned in the background art through the following scheme.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent mobile visitation route optimization method for end-user business personnel, comprising the following steps: S1: Real-time collection of target customer status data, salesperson's current execution dynamic data, and external environment change data, recorded as target trip decision data, and identification of abnormal dynamic data of business based on target trip decision data and preset rules; S2: Based on dynamic data of business anomalies, calculate the comprehensive priority score of each customer, sort them by score, and plan the initial optimal visit route by combining the salesperson's real-time location and road conditions. S3: Continuously monitor target trip decision data, and when an anomaly is detected, update the customer's overall priority score in real time. Based on the updated customer's overall priority score, identify the current highest value customer, recalculate and push optimized routes. S4: Push optimized routes to sales staff terminals and mark customers with low inventory, tight deadlines, or high value; sales staff submit dynamic execution data in real time, and the system periodically adjusts customer priority scoring rules and dynamic route adjustment strategies.

[0007] Preferably, the customer status data includes customer location, inventory level, order amount, and scheduled visit time period; The salesperson's current dynamic data includes the salesperson's real-time location, dwell time, average speed, completed tasks, remaining tasks, and remaining time. The external environmental change data includes road conditions and weather.

[0008] Preferably, the preset rules include: Inventory alert rule: When a customer's inventory level is lower than 50% of the safety stock level and the weekly consumption rate exceeds 40% of the historical average, an inventory emergency is triggered. Salesperson execution rules: When a salesperson spends more than 1.5 times the preset visit duration at a single customer's location, and the number of visits the salesperson needs to complete per hour exceeds 1.3 times the number of visits they have completed per hour in the past, an execution efficiency anomaly is triggered. External environment rules: When a salesperson's average speed on the target road segment is less than 40% of the historical average speed for the same period on that road segment, and this state is maintained for 10 consecutive minutes, or when the meteorological department issues an orange or higher level warning for heavy rain, heavy snow, etc., an abnormal traffic risk is triggered. Time conflict rule: When two or more customer appointment time slots overlap by more than 30 minutes, or when the shortest travel time between customers is greater than the interval between time slots, a time slot conflict exception is triggered.

[0009] Preferably, the dynamic data of business anomalies includes: Inventory alert anomaly data: includes the customer ID that triggered the inventory alert rule, current inventory level, safety stock level, and weekly average consumption rate; Execution efficiency anomaly data includes the salesperson identifier that triggered the salesperson execution rule, the duration of stay at the current customer, the preset visit duration for that customer, the remaining task load, the remaining available time, and the historical average task efficiency; Traffic risk anomaly data: includes road segment identification that triggers external environmental rules, the average speed of salespersons on that road segment, the average speed of that road segment during the same period in the past 30 days, the duration of the abnormal state, and the type and level of weather warnings; Time-time conflict anomaly data: includes the customer identifier that triggered the time conflict rule, their respective appointment time slots, the shortest travel time between customers, and the duration of time-time overlap.

[0010] Preferably, the comprehensive priority score is obtained by filtering abnormal data and generating four types of abnormal indicator vectors: inventory shortage vector, execution efficiency abnormal vector, passage risk vector, and time period conflict vector, which are standardized into coefficient vectors X in the range of 0-1; based on the preset weight column vector W, the comprehensive priority score is obtained through inner product operation.

[0011] Preferably, the highest value customers exclude those who have been visited or those who cannot be received, and simultaneously meet the following conditions: the profit ratio in the past 6 months is ≥5%, there is business abnormality, and the probability of failure to close a deal due to the abnormality is 20% or more higher than normal. If the first two conditions are met but the priority is not in the top 5%, the customer is adjusted according to the profit excess ratio. If the customer is in the top 10% after adjustment, the customer is considered. Otherwise, the customer is pushed to the management terminal for confirmation.

[0012] Preferably, the customer priority scoring rules include: Quarterly weight adjustment: The weights of each indicator are adjusted quarterly based on the correlation between the previous quarter and the actual transaction rate. Indicators with a high correlation have a 15% increase in weight, while indicators with a low correlation have an 8% decrease in weight. Temporary weighting of anomalies: When a customer generates abnormal dynamic data in their business, the weights of the four types of anomaly indicator vectors corresponding to that abnormal dynamic data are temporarily increased by 30% within 24 hours. New Customer Support: For new customers with a cooperation period of less than 3 months, when calculating the cooperation stability value, the frequency of cooperation in the past year will be replaced by the order growth rate in the past 2 months, and the weight of the cooperation stability value will be temporarily increased by 15%. Negative record decay: The impact of past customer complaints decreases over time. The first month is calculated as 100% impact, the second to third month as 50% impact, the fourth to sixth month as 20% impact, and the impact is no longer calculated after 6 months.

[0013] Preferably, the route dynamic adjustment strategy includes: Risk avoidance strategy: When there is an abnormal risk of traffic in the route, calculate the benefit value of each detour route and select the detour route with the highest benefit value and greater than 0.5; Customer clustering strategy: Divide customers to be visited into regions based on geographical location, and sort customers in the same region from high to low based on their comprehensive priority score, and plan them into a continuous visit sequence; Priority mutation response strategy: When a customer's overall priority score increases by more than 25% within 15 minutes, immediately insert them into the current route.

[0014] The technical effects and advantages of this invention are as follows: 1. This invention collects dynamic data on four types of business anomalies in real time: low inventory, execution efficiency, traffic risks, and time period conflicts. It then transforms these data into standardized indicators and conducts a comprehensive evaluation, enabling customers to prioritize visit urgency based on data rather than making vague judgments. This helps sales staff to prioritize high-urgency needs and reduce losses caused by decision-making delays. 2. The indicators of each dimension of this invention are weighted according to a basic ratio, and the weights are dynamically adjusted in special circumstances to ensure that resources are tilted towards core needs, reduce ineffective work, and improve execution efficiency; 3. This invention avoids external interference in advance by assessing road conditions and weather warnings; it also scientifically resolves time conflicts, reduces complaint rates, and improves business stability by analyzing the overlap of reservation time slots and travel time. 4. This invention adopts a modular design to support independent optimization or the addition of new dimensions, allowing for flexible adjustments; the linear evaluation logic makes it easy to trace the source of scores, facilitating understanding and operation for business personnel, and enabling rapid adaptation to different scenarios with strong feasibility. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall structure of the present invention; Figure 2 This is a schematic diagram of the business anomaly dynamic data identification process structure of the present invention; Figure 3 This is a schematic diagram of the overall priority and initial route planning flowchart structure of the present invention; Figure 4 This is a schematic diagram of the dynamic monitoring and route optimization flowchart of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] like Figure 1-4 The intelligent mobile visit schedule optimization method for end-user business personnel, as shown, includes the following steps: S1: Real-time collection of target customer status data, salesperson's current execution dynamic data, and external environment change data, recorded as target trip decision data, and identification of abnormal dynamic data of business based on target trip decision data and preset rules; Target customer status data is collected in real time via inventory sensors with IoT modules deployed on customer terminals. This data includes current inventory levels, safety stock levels, payments received in the past 6 months, actual visit frequency in the past 6 months, and scheduled visit times. Customer location information is collected when a customer first enters the system. Salespeople or customers submit detailed customer addresses via their terminals. The system then calls a map service interface to convert the text address into precise latitude and longitude coordinates, which are then bound to the customer's identifier and stored.

[0018] It should be noted that inventory data is automatically uploaded every 15 minutes via sensors; the amount received and the number of visits are automatically synchronized through the interaction records between the financial system and the salesperson's terminal, and the data of the previous day is summarized and updated at midnight every day; the appointment time slot is manually entered by the customer through the system backend, and then synchronized to the central database in real time.

[0019] It is important to note the safety stock level. , where D avg Let L be the average daily consumption, L be the replenishment lead time, and F be the volatility coefficient. The replenishment lead time is the average time from when the salesperson submits an order to when the goods are delivered to the customer. The volatility coefficient is determined based on the customer's weekly outbound volume fluctuations over the past three months: where the fluctuation range... W max This represents the largest weekly outflow in nearly three months. min This is the lowest weekly outbound volume in nearly three months, W avg This represents the average weekly outbound volume over the past three months. When F A If it exceeds 30%, F is taken as 0.3. A When the percentage is between 10% and 30%, F is taken as 0.2. A When it is below 10%, F is taken as 0.1.

[0020] The salesperson's current dynamic data is collected in real time via their handheld smart terminal. This data includes the salesperson's location, total revenue received in the past 6 months, average number of visits per month in the past 6 months, planned arrival time, duration of stay at the current customer's location, remaining tasks, and remaining available time. Location information and duration of stay are updated every 30 seconds via GPS positioning. Upon arrival at the customer's location, scanning the customer's unique QR code triggers the visit timer, and scanning the QR code upon departure ends the timer. Remaining tasks and remaining available time are automatically calculated by the system based on the daily task list and preset work periods, and are refreshed every 5 minutes.

[0021] External environmental change data is obtained in real time through the meteorological bureau's API, including road segment identification, real-time traffic conditions, average speed over the past 30 days, duration of abnormal conditions, and type and level of weather warnings for the salesperson's current location and planned route. Traffic condition data includes average speed and congestion level for each road segment, updated every 2 minutes. Weather warning data is synchronized to the system within 10 minutes of a warning being issued.

[0022] The preset rules include: Inventory alert rules: When a customer's inventory level is lower than 50% of the safety stock level and the weekly consumption rate exceeds the historical average of 40%, an inventory emergency is triggered; the customer identifier, current inventory level, safety stock level, and weekly average consumption rate are recorded. Salesperson execution rules: When a salesperson's stay at a single customer exceeds 1.5 times the customer's preset visit duration, and the number of visit tasks that the current salesperson needs to complete per hour exceeds 1.3 times the number of visit tasks completed per hour in the past, an execution efficiency anomaly is triggered; the salesperson's identifier, current customer stay duration, preset visit duration, remaining task load, remaining available time, and historical average task efficiency are recorded. External environment rules: When a salesperson's average speed on a target road segment is lower than 40% of the historical average speed for the same period on that road segment, and remains in this state for 10 consecutive minutes, or when the meteorological department issues an orange or higher level warning for heavy rain, heavy snow, etc., an abnormal traffic risk is triggered; record the road segment sign, real-time speed, historical speed for the same period, duration of the abnormality, and meteorological warning information; Time conflict rules: When two or more customer appointment time slots overlap for more than 30 minutes, or the shortest travel time between customers is greater than the interval between time slots, a time slot conflict exception is triggered; record the customer ID, appointment time slot, shortest travel time, and overlap duration.

[0023] The system compares the collected target travel data with preset rules, automatically filters out abnormal data that meets the conditions, and stores it in the business anomaly database after automatic cleaning by the system.

[0024] S2: Based on dynamic data of business anomalies, calculate the comprehensive priority score of each customer, sort them by score, and plan the initial optimal visit route by combining the salesperson's real-time location and road conditions. The comprehensive priority score filters abnormal data through preset rules, generating four types of abnormal indicator vectors: inventory shortage vector, execution efficiency abnormal vector, passage risk vector, and time period conflict vector. These vectors are combined into an abnormal data matrix A with a dimension of m*4, where m is the number of abnormal customers and the column vectors correspond to the four types of abnormal indicators.

[0025] Transform the four types of outlier vectors into standardized coefficient vectors in the 0-1 interval. Where x1 is the inventory shortage coefficient, x2 is the execution efficiency coefficient, x3 is the passage risk coefficient, and x4 is the time period conflict coefficient.

[0026] It should be noted that, When the weekly consumption rate exceeds 5 times the safety stock level, multiply by the adjustment matrix by 1.2. Here, ISR represents the inventory gap percentage, and its analysis formula is... SS represents the safety stock level, and I represents the current stock level (when the current stock level is greater than the safety stock level, the gap ratio is 0). It should be noted that, When the task pressure coefficient is greater than 1.5 times the historical average task efficiency, it is multiplied by the adjustment matrix by 1.1 times; where TSR is the dwell time percentage, and its analysis formula is as follows: T a T represents the actual duration of stay. p The actual reservation duration (when the actual stay exceeds the actual reservation duration or the actual stay is 0 due to the customer's temporary inability to be received, the stay duration percentage is taken as 1). It should be noted that the analytical formula for the task stress coefficient is as follows: W r T represents the remaining number of tasks (unit: items). r Remaining available time (in hours).

[0027] It should be noted that, When the duration of the abnormal condition exceeds 0.5 hours or the weather warning level is red, multiply by the adjustment matrix by 1.3, where VR is the speed deviation rate, and its analysis formula is as follows: V c V represents the salesperson's current average speed. h VR is the average speed of the salesperson over the past 30 days (VR is 1 when the current average speed of the salesperson is higher than the historical average, and VR is 0.1 when the current average speed of the salesperson is 0).

[0028] It should be noted that, When the shortest travel time between customers is greater than 1.5 times the interval between appointment slots, multiply by a 1.2-fold adjustment matrix, where TR is the time slot overlap rate, and its analysis formula is as follows: T o T represents the overlap between two client appointment times. t The total duration of the scheduled visit times for the two clients (if the time slots completely overlap, the overlap rate is 1).

[0029] Set weight column vector Customer overall priority score ,Right now The system sorts all customers’ S-values ​​in descending order and generates a priority vector, where the elements are customer identifiers and corresponding scores.

[0030] It should be noted that the preset weight column vector The assignment is based on the following: The weighting was primarily based on "the degree of impact of abnormal indicators on terminal business conversion rate and customer retention," and was derived from the company's historical field business data over the past 12 months, including statistics from 2360 valid visit cases. Inventory shortage anomaly: Historical data shows that "supply disruption complaints" caused by failure to visit customers in a timely manner due to inventory shortages account for 42% of the total complaints. Moreover, the customer repurchase rate increases by 38% after such anomalies are resolved. This is a key factor affecting "core business continuity" and is therefore given the highest weight.

[0031] Abnormal execution efficiency: Low execution efficiency of salespersons (such as excessive stay of a single customer or backlog of tasks) will lead to a 25% decrease in the average daily effective visits, indirectly affecting the appointment fulfillment rate of 3-5 customers, and has a significant impact on "overall visit productivity", second only to inventory shortage in weight.

[0032] Traffic risk anomalies: Although such anomalies may delay travel, 80% of the negative impact can be avoided through strategies such as detours and time adjustments. Historical data shows that only 9% of uncompleted transactions are due to traffic risks, and the impact on core business is weaker than the previous two types of indicators.

[0033] Time slot conflict anomaly: Time slot conflicts can be resolved by negotiating with the customer to adjust the appointment time, and only 12% of conflicts will lead to no-shows. Its interference with the "transaction result" is close to that of traffic risk, so it is given the same weight as traffic risk.

[0034] Given the characteristics of fast-moving consumer goods and retail businesses, such as "rapid inventory turnover and high visit frequency," inventory shortages are directly related to the risk of sales interruption for customers and must be addressed first. In contrast, traffic risks and time conflicts are considered "coordinable interferences" and can be dealt with later when resources are limited. The weighting of these issues aligns with the core requirements of field operations in the industry.

[0035] The initial weights of 0.4, 0.3, 0.15, and 0.15 are the "baseline weights," which will be iteratively optimized through a quarterly weight adjustment mechanism. If the statistics of a certain quarter show that the proportion of uncompleted transactions caused by time period conflicts rises to 18%, more than double the original 9%, then the weight of the time period conflict vector will be increased from 0.15 to 0.17 according to the rule of increasing the weight of indicators with high correlation by 15%.

[0036] It should be noted that the initial optimal visit route is planned with the salesperson's current location as the starting vector P0, the customer location matrix as P (n*2, where n is the number of customers, and the column vectors are latitude and longitude), and the distance matrix D (n*n) is calculated, where D... ij Let be the straight-line distance between customers i and j. Combining the priority vector and the real-time traffic coefficient matrix R (n*n, elements being road segment congestion coefficients), an improved Dijkstra algorithm is used to solve for the optimal path matrix: Path cost d represents the actual transportation distance (non-straight-line distance) between the salesperson's current location and the location of the next target customer, C f W represents the congestion coefficient of a road segment. p This indicates the priority weight; the path cost calculation constraint is that the arrival time falls within the vector interval of the customer's scheduled time slot. The final generated initial route is stored in the form of an ordered vector, containing the customer visit order and the expected arrival timestamp.

[0037] It should be noted that the actual transportation route distance is based on the salesperson's current location vector (GPS coordinates (x-axis)). t y t and the target customer location vector P i Call the map service interface to obtain the optimal driving route between two points that complies with traffic rules, and calculate the physical length of the route (in kilometers); if it is a series of visits to multiple customers, the distance is the location vector P of customer A. A Vector P to customer B's location B The actual route distance; the distance data is updated in real time with the traffic condition matrix. When a road segment is closed, the system automatically switches to the alternative route and recalculates the distance.

[0038] It should be noted that the road congestion coefficient V f V represents the free-flow velocity. c The real-time average speed is represented by the free-flow speed, which is a preset baseline value based on the road segment type and stored in the road condition basic parameter matrix; the real-time average speed is taken from the road segment speed collected from external environmental data and updated every 2 minutes.

[0039] It should be noted that the priority weight is based on the calculated overall customer priority score. For all customers currently to be visited, the maximum value S of their overall priority scores is taken. max , .

[0040] It should be noted that, based on the analysis of abnormal customers, non-abnormal customer data is added to form a full customer priority matrix B with n*5 dimensions. The first 4 dimensions are the same as matrix A, and the 5th dimension is the abnormality indicator (0 for non-abnormal and 1 for abnormal).

[0041] S3: Continuously monitor target trip decision data, and when an anomaly is detected, update the customer's overall priority score in real time. Based on the updated customer's overall priority score, identify the current highest value customer, recalculate and push optimized routes. The system is equipped with a monitoring module. The customer status matrix is ​​refreshed every 30 minutes, and the inventory coefficient vector x1 is recalculated in real time according to changes in inventory. The salesperson execution matrix is ​​refreshed every 10 minutes, and the location vector and remaining task vector trigger dynamic adjustments (x2). The road condition matrix iterates every 2 minutes, and changes in the speed deviation rate cause the x3 vector to be updated. It is immediately recalculated when a weather warning is triggered. Using the most recent value as the baseline, if the change rate of any element in any parameter matrix exceeds 10%, the system automatically initiates the vector update process and re-pushes the optimized route.

[0042] It should be noted that the vector update recalculates the X vector based on the latest data, generates a new priority vector through the inner product S=X*W, and locks the highest-scoring customer as the core node; using the current position vector P... t Starting from the root, the new priority vector serves as the weight. Vectors of customers whose visits have been completed are removed. Customers whose priority drops by 50% are assigned lower weights in the path matrix. When adding a high-priority customer, it is inserted into the nearest matrix node. The optimized route matrix includes the adjusted visit order and time window, and when pushed via the terminal app, it includes an explanation of the vector changes.

[0043] It should be noted that the highest value customers must simultaneously meet the following criteria: their overall priority score ranks in the top 5% of all customers to be visited; there is at least one abnormal business dynamic data point, and the probability of not closing the deal due to this abnormality is 20% or more higher than the probability of not closing the deal under normal circumstances; the salesperson arrives at the customer from the current location earlier than the latest acceptable visit deadline of the customer; and the profit generated from cooperation with the customer in the past 6 months accounts for 5% or more of the salesperson's total profit in the past 6 months.

[0044] S4: Push optimized routes to sales staff terminals and mark customers with low inventory, tight deadlines, or high value; sales staff submit dynamic execution data in real time, and the system periodically adjusts customer priority scoring rules and dynamic route adjustment strategies.

[0045] The system converts the route matrix into a visual path, using color to indicate vector weights: red for x1, yellow for x2, blue for x3, and green for x4, making it easy for sales staff to intuitively understand the impact of each dimension. When submitting execution results via the app, sales staff can click "Complete Visit" to update the task status vector, fill in anomaly feedback to correct time period conflict vectors, and upload actual travel time for calibrating the road condition coefficient matrix. All feedback data is stored in a diagonal matrix format, accurately recording the error values ​​of each vector dimension. It's important to note that the system initiates a closed-loop optimization process monthly for weight matrix and algorithm parameter optimization: During the weight matrix W update phase, the covariance between each vector dimension and the transaction rate is calculated. If the correlation between x1 and the transaction rate is greater than 0.6, the weight is increased by 15%; if the correlation is less than 0.2, the weight is decreased by 8%, thus forming a new weight matrix W. In the adjustment coefficient optimization phase, the adjustment matrix is ​​corrected based on error values ​​in the feedback data. For example, the 1.2 multiplier for low inventory levels will be adjusted to 1.3 based on the actual supply disruption rate. The updated matrix parameters are automatically loaded into the system for vector operations in the next cycle, thus achieving a complete closed loop of data collection, matrix operations, execution feedback, and parameter iteration.

[0046] It should be noted that the system extracts historical data from the past 30 days each month to construct two vectors: one is a sequence of vector dimension values ​​for each customer, and the other is a sequence of conversion rates for the corresponding customer. The correlation between the two is calculated using the covariance formula. The conversion rate, C... a C represents the actual number of customers who made a purchase. t This represents the total number of customers visited.

[0047] It should be noted that, , where x ik Let be the value of the i-th dimension of the vector on day k. Let y be the mean of the vector dimension. k Let be the transaction rate on day k. is the average transaction rate, and n is the sample size (30 days).

[0048] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other. In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the intelligent mobile visit itinerary for end-user business personnel, characterized in that, include: S1: Collect target trip decision data in real time, and identify abnormal dynamic data of business based on target trip decision data and preset rules; The target trip decision data includes: target customer status data, salesperson's current execution dynamic data, and external environment change data; S2: Based on dynamic data of business anomalies, calculate the comprehensive priority score of each customer, sort them from high to low scores, and plan the initial optimal visit route by combining the salesperson's real-time location and road conditions. S3: Continuously monitor target trip decision data, and when an anomaly is detected, update the customer's overall priority score in real time. Based on the updated customer's overall priority score, identify the current highest value customer, recalculate and push optimized routes. S4: Push the optimized route to the salesperson's terminal and mark customers who are running out of inventory, have tight deadlines, or are high-value customers; Sales representatives submit real-time execution data, and the system periodically adjusts customer priority scoring rules and route dynamic adjustment strategies.

2. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The customer status data mentioned in S1 includes customer location, inventory level, order amount, and scheduled visit time period; The salesperson's current dynamic data includes the salesperson's real-time location, dwell time, average speed, completed tasks, remaining tasks, and remaining time. The external environmental change data includes road conditions and weather.

3. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The preset rules mentioned in S1 include: Inventory alert rule: When a customer's inventory level is lower than 50% of the safety stock level and the weekly consumption rate exceeds 40% of the historical average, an inventory emergency is triggered. Salesperson execution rules: When a salesperson spends more than 1.5 times the preset visit duration at a single customer's location, and the number of visits the salesperson needs to complete per hour exceeds 1.3 times the number of visits they have completed per hour in the past, an execution efficiency anomaly is triggered. External environment rules: When a salesperson's average speed on the target road segment is less than 40% of the historical average speed for the same period on that road segment, and this condition is maintained for 10 consecutive minutes, or when the meteorological department issues an orange or higher level warning, an abnormal traffic risk is triggered. Time conflict rule: When two or more customer appointment time slots overlap by more than 30 minutes, or when the shortest travel time between customers is greater than the interval between time slots, a time slot conflict exception is triggered.

4. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The dynamic data of business anomalies mentioned in S1 includes: Inventory alert anomaly data: includes the customer ID that triggered the inventory alert rule, current inventory level, safety stock level, and weekly average consumption rate; Execution efficiency anomaly data includes the salesperson identifier that triggered the salesperson execution rule, the duration of stay at the current customer, the preset visit duration for that customer, the remaining task load, the remaining available time, and the historical average task efficiency; Traffic risk anomaly data: includes road segment identification that triggers external environmental rules, the average speed of salespersons on that road segment, the average speed of that road segment during the same period in the past 30 days, the duration of the abnormal state, and the type and level of weather warnings; Time-time conflict anomaly data: includes the customer identifier that triggered the time conflict rule, their respective appointment time slots, the shortest travel time between customers, and the duration of time-time overlap.

5. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The comprehensive priority score mentioned in S2 generates four types of abnormal indicator vectors—inventory shortage vector, execution efficiency abnormal vector, passage risk vector, and time period conflict vector—by filtering abnormal data and standardizing them into a coefficient vector X in the range of 0-1. Based on the preset weighted column vector W, the comprehensive priority score is obtained through inner product operation.

6. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The highest value customers mentioned in S3 exclude customers who have been visited or who cannot be received, and at the same time meet the following conditions: the profit ratio in the past 6 months is ≥5%, there is business abnormality, and the probability of failure to close the deal is 20% or more higher than normal. If the first two conditions are met but the priority is not in the top 5%, the weighted adjustment is made according to the profit excess ratio. If the adjusted customer is in the top 10%, it is considered as such. Otherwise, the request is pushed to the management terminal for confirmation.

7. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The customer priority scoring rules described in S4 include: Quarterly weight adjustment: The weights of each indicator are adjusted quarterly based on the correlation between the previous quarter and the actual transaction rate. Indicators with a high correlation have a 15% increase in weight, while indicators with a low correlation have an 8% decrease in weight. The degree of correlation is determined by the correlation coefficient between the indicator and the actual transaction rate. A correlation coefficient greater than 0.6 indicates a high degree of correlation, while a correlation coefficient less than 0.2 indicates a low degree of correlation. Temporary weighting of anomalies: When a customer generates abnormal dynamic data in their business, the weights of the four types of anomaly indicator vectors corresponding to that abnormal dynamic data are temporarily increased by 30% within 24 hours. New Customer Support: For new customers with a cooperation period of less than 3 months, when calculating the cooperation stability value, the frequency of cooperation in the past year will be replaced by the order growth rate in the past 2 months, and the weight of the cooperation stability value will be temporarily increased by 15%. Negative record decay: The impact of past customer complaints decreases over time. The first month is calculated as 100% impact, the second to third month as 50% impact, the fourth to sixth month as 20% impact, and the impact is no longer calculated after 6 months.

8. The intelligent mobile visitation optimization method for terminal business personnel according to claim 1, characterized in that: The route dynamic adjustment strategy described in S4 includes: Risk avoidance strategy: When there is an abnormal traffic risk in the route, calculate the benefit value of each detour route and select the detour route with the highest benefit value and greater than 0.5; the benefit value is the product of the time saved by the detour and the customer priority weight, minus the product of the increased distance by the detour and the unit distance cost. Customer clustering strategy: Divide customers to be visited into regions based on geographical location, and sort customers in the same region from high to low based on their comprehensive priority score, and plan them into a continuous visit sequence; Priority mutation response strategy: When a customer's overall priority score increases by more than 25% within 15 minutes, immediately insert them into the current route.

Images