Intelligent locker distribution method in shared office scene

Abstract

The invention discloses an intelligent locker distribution method in a shared office scene, and belongs to the field of locker application, and the method comprises the following steps: collecting data, and carrying out the preprocessing of the data; carrying out feature modeling; performing dynamic threshold adjustment and resource optimization; and intelligent distribution of storage cabinets is realized. According to the method, the utilization rate condition of the storage cabinets can be monitored and analyzed in real time, the types (sharing / fixing) and the number of the storage cabinets are dynamically adjusted according to actual requirements, and the problem that resource shortage occurs due to the fact that the utilization rate of the storage cabinets is too high or resource waste is caused due to the fact that the utilization rate of the storage cabinets is too low is solved. The types and the number of the storage cabinets are automatically adjusted by setting a utilization rate threshold and adopting a certain logic and algorithm, so that the utilization efficiency of the storage space is improved, and the existence rate of zombie cabinets is reduced; and the workload of manual management and maintenance can be reduced, the operation cost is reduced, the user experience of the locker system is improved, and the waiting time is shortened, so that the use satisfaction degree of the locker is improved.

Description

Technical Field

[0001] The present invention belongs to the application field of lockers, and particularly relates to a method for allocating intelligent lockers in a shared office scenario. Background Art

[0002] With the development of the sharing economy, the shared office scenario has become increasingly popular. In the shared office scenario, intelligent lockers are usually equipped for employees to temporarily store personal items. However, the existing intelligent lockers have the problem of low utilization rate. On the one hand, traditional lockers are divided into fixed types, such as 70% fixed lockers + 30% shared lockers, and the ratio cannot be dynamically adjusted according to actual needs, resulting in over - utilization of some lockers and resource shortage, while under - utilization of some lockers and resource waste. On the other hand, there is a lack of real - time monitoring of the usage rate of lockers and historical data analysis, and it is impossible to identify "zombie lockers" (lockers that are idle for a long time). In addition, the existing shared lockers adopt the "first - come, first - served" allocation mode, and do not optimize the allocation path by combining user preferences, usage habits and spatial location, resulting in poor user experience.

[0003] The root cause of the above problems is that the existing intelligent locker system lacks a real - time monitoring and analysis mechanism for the utilization rate of lockers, and cannot dynamically adjust the type and quantity of lockers according to actual needs. This not only reduces the utilization rate of storage space, but also increases the management and maintenance costs.

[0004] The existing system does not establish a closed - loop feedback mechanism of "data collection → dynamic prediction → intelligent allocation", and lacks the adaptive ability to user behavior and demand fluctuations; therefore, there is an urgent need for a method for allocating intelligent lockers. Summary of the Invention

[0005] In order to make up for the deficiencies of the prior art, the purpose of the present invention is to provide a method for allocating intelligent lockers in a shared office scenario, which can calculate and analyze the usage rate of each locker based on the utilization rate of lockers within a period of time, set a certain threshold, automatically adjust the type (shared / fixed) of lockers, or add or reduce lockers, through a certain logic and algorithm, so as to greatly improve the utilization rate of storage space and reduce the existence rate of zombie lockers.

[0006] The technical problems solved by the present invention can be realized by the following specific technical solutions:

[0007] The described method for allocating intelligent lockers in a shared office scenario includes the following steps:

[0008] Step 1: Collect data and pre - process the data.

[0009] Step 2: Feature modeling;

[0010] Step 3: Dynamic threshold adjustment and resource optimization;

[0011] Step 4, intelligent allocation of lockers.

[0012] Further, in the said Step 1, the collected data includes user portraits, environmental data, and cabinet status information.

[0013] Further, in the said Step 1, the preprocessing of the data includes data cleaning, data standardization, and data encoding.

[0014] Further, in the said Step 2, the feature modeling is to construct an association matrix, specifically as follows:

[0015] Step 2.1, initialize the association matrix: If there are m users and n lockers in total, then the association matrix A is an m×n matrix, and all elements in the matrix are 0 initially;

[0016] Step 2.2, fill the association matrix: Traverse the usage records of each user. For the locker j used by user i, update the element A in the association matrix ij ; where, if user i has only used locker j once, then A ij = 1; if user i has used locker j multiple times, then the value of A ij is the number of usage times;

[0017] Step 2.3, normalize the association matrix: Divide each element by the total number of times the user has used the locker.

[0018] Further, in the said Step 3, the threshold judgment time range is T1, and the server automatically calculates the average usage rate of T1. The average usage rate = total usage duration / total available duration × 100%; Compare the average usage rate of each locker with the preset upper and lower threshold values during peak hours. If it is lower than the lower threshold value, the locker type is converted. If it is higher than the upper threshold value, the locker type remains unchanged; If 1 / 2 of the lockers in the same locker grid are continuously higher than the average usage rate 3 times during peak hours, the system triggers a message reminder to the management staff, and new lockers of the same type are added in the same area.

[0019] Further, if the average usage rate is within the peak threshold range, then judge the off-peak hours. If it is lower than the lower limit value during off-peak hours, the locker type is adjusted, that is, it is fixed to be converted to shared, and vice versa. If it is higher than the upper limit value within the threshold range during off-peak hours, the locker type remains unchanged; If the usage rate of the cabinet is < 5% for N consecutive days and there is no weight change, it is marked as a "zombie cabinet" and triggers recycling or reallocation.

[0020] Further, in step 4, the intelligent allocation includes fixed cabinet allocation and shared cabinet allocation. The fixed cabinet allocation method is based on a priority queue and an elastic recycling mechanism. The priority queue is sorted according to the historical usage rate of users, and the evaluation of usage efficiency is introduced. If the usage rate of a user is lower than the threshold within a specified time, it will be automatically adjusted to a shared cabinet and the user will be notified. The shared cabinet allocation is based on a multi-objective optimization model, subject to the matching degree of cabinet types and the constraint of limited reservation occupancy of cabinets.

[0021] Further, the specific process of the fixed cabinet allocation is as follows:

[0022] ① Construct a priority queue: Collect the historical usage data of all users, calculate the historical usage rate of each user according to the formula "average daily usage duration × frequency", sort all users from high to low according to the historical usage rate, and construct a priority queue;

[0023] ② Allocate fixed cabinets: Traverse the priority queue. For each high-priority user, check if there is an available fixed cabinet. If there is an available fixed cabinet, allocate the fixed cabinet to the user and record the corresponding allocation information;

[0024] ③ Usage efficiency evaluation and elastic recycling: Regularly check the usage situation of users with allocated fixed cabinets, calculate the average usage rate of each user in the past 3 days. If it is lower than the preset threshold, automatically adjust the fixed cabinet of this user to a shared cabinet. At the same time, the system sends a notice to this user, informing that the type of the locker has been changed to a shared cabinet.

[0025] Further, the specific process of the shared cabinet allocation is as follows:

[0026] ① Data collection: Collect the preference label data of all users, assign corresponding scores to each preference label, determine the workstation location of each user and the location of the lockers, for calculating the distance from the workstation to the target locker;

[0027] ② Calculate the user satisfaction score: For each user, calculate the sum of the preference label scores according to their preference labels;

[0028] ③ Calculate the distance of the pick-up path: Based on the indoor navigation path algorithm, calculate the distance from each user's workstation to each shared cabinet;

[0029] ④ Solve the multi-objective optimization model: Construct a multi-objective optimization model, and the objective function is to maximize the user satisfaction score and minimize the distance of the pick-up path;

[0030] ⑤ Allocate shared cabinets: Traverse all shared cabinets. For each shared cabinet, allocate it to the user with the highest comprehensive score.

[0031] Compared with the prior art, the present invention has the following advantages:

[0032] (1) The method of the present invention can monitor and analyze the utilization rate of the lockers in real time, dynamically adjust the types (shared / fixed) and quantities of the lockers according to actual needs, and avoid the problems of resource shortage due to excessive utilization rate of the lockers or resource waste due to too low utilization rate of the lockers.

[0033] (2) By setting a utilization rate threshold and adopting certain logics and algorithms, the present invention automatically adjusts the types and quantities of the lockers, which not only improves the utilization efficiency of the storage space and reduces the existence rate of zombie lockers, but also can reduce the workload of manual management and maintenance, lower the operation cost, improve the user experience of the locker system, reduce the waiting time, and thus enhance the satisfaction of using the lockers.

[0034] (3) The present invention supports data interaction with the office space management system, such as the workstation reservation platform, the meeting reservation platform, etc., to achieve collaborative resource scheduling. Description of the Drawings

[0035] Figure 1 is the flowchart for constructing data collection and feature modeling of the present invention;

[0036] Figure 2 is the flowchart for dynamic threshold adjustment and resource optimization of the present invention;

[0037] Figure 3 is the schematic diagram of the intelligent allocation algorithm process of the present invention. Detailed Embodiment

[0038] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0039] The present invention is directed to an allocation method for intelligent lockers. The intelligent lockers are prior art and include an NFC / face recognition module, an electronic lock, an integrated weight sensor, a mobile phone APP, a background management system, etc.

[0040] An intelligent locker allocation method in a shared office scenario includes the following contents:

[0041] Step 1: Collect data and preprocess the data.

[0042] Such as Figure 1As shown in the figure, data collection mainly collects information such as user portraits, environmental data, and cabinet status. User portraits rely on users' registration information and behavioral data. Registration information includes basic information such as gender, height, department, and work station location; behavioral data refers to the frequency of users' use of lockers and their own storage preference notes, such as being close to the reserved work station, near the elevator, and often on business trips; cabinet status refers to the occupancy duration, number of openings and closings, and changes in item weight of each locker.

[0043] Collect the above data and preprocess the data, then construct a user-cabinet association matrix, and extract key features such as "user usage stability index", "cabinet area heat value", and "department with the greatest storage demand".

[0044] Among them, data preprocessing includes data cleaning, data standardization, and data encoding.

[0045] (1) Data cleaning.

[0046] ① Process missing values. For missing values in user registration information, such as gender, height, department, etc., if the missing rate is low, the mean, median, or mode can be used for filling; if the missing rate is high, consider deleting the record.

[0047] ② For missing values in behavioral data, such as usage frequency and storage preferences, they can be filled according to the information of similar users.

[0048] ③ Process outliers. For outliers in cabinet status data, such as too long occupancy duration and too many opening and closing times, statistical methods (such as the Z-score method) can be used for identification and processing. If the absolute value of the Z-score of a data point is greater than 3, then the data point is considered an outlier and can be deleted or corrected.

[0049] (2) Data standardization. Standardize the collected numerical data to make it have the same scale. Commonly used standardization methods include Min-Max standardization and Z-score standardization.

[0050] (3) Data encoding. For categorical data, such as gender, department, etc., encoding processing is required to convert it into numerical data. Commonly used encoding methods include one-hot encoding and label encoding. One-hot encoding converts each category into a binary vector, where only one element is 1 and the rest are 0; label encoding assigns a unique integer label to each category.

[0051] Step 2: Feature modeling (i.e., constructing an association matrix).

[0052] Step 1: Initialize the association matrix. Assume there are m users and n lockers in total. Then the association matrix A is an m×n matrix, and all elements in the matrix are initially 0.

[0053] Step 2: Populate the association matrix. Traverse the usage records of each user. For locker j used by user i, update the element A in the association matrix. ij ; where, if user i has used locker j only once, then A ij = 1; if user i has used locker j multiple times, then the value of A ij is the number of usage times.

[0054] Step 3: Normalize the association matrix. To make the elements in the association matrix comparable, perform normalization processing on it. A common normalization method is to divide each element by the total number of times the user uses the lockers.

[0055] Step 3: Dynamic threshold adjustment and resource optimization.

[0056] As Figure 2 shown, if the threshold judgment time range is (T1), the server automatically calculates the average usage rate of (T1). The calculation formula is: average usage rate = total usage duration / total available duration × 100%. Compare the average usage rate of each locker with the preset upper threshold (such as 80%) and lower threshold (such as 30%) during peak hours. If it is lower than the lower threshold, the locker type is converted. If it is higher than the upper threshold, the locker type remains unchanged. If 1 / 2 of the lockers in the same locker grid are continuously higher than the average usage rate 3 times during peak hours, the system triggers a message reminder to the management staff, and new lockers of the same type are added in the same area.

[0057] In addition, during peak hours, if the average usage rate is within the threshold range (30% - 80%), the off-peak hours are judged. If it is lower than the lower limit during off-peak hours, the locker type is adjusted, that is, it is fixed to be converted to shared, and vice versa; if it is within the threshold range or higher than the upper limit during off-peak hours, the locker type remains unchanged (if the off-peak hour threshold range is 20% - 40%, when the average usage rate during off-peak hours is below 20%, the locker type is adjusted to shared; when it is between 20% - 40% or higher than 40%, the locker type remains unchanged.); if the usage rate of the cabinet is < 5% for N consecutive days and there is no weight change, it is marked as a "zombie cabinet" and triggers recycling or reallocation.

[0058] Note: The threshold is only an example. In actual applications, it needs to be adjusted and optimized according to the specific usage data and business requirements of smart lockers in the shared office scenario to achieve the best resource allocation and usage effect.

[0059] Step 4: Intelligent allocation of lockers.

[0060] Based on the data statistics and analysis module, the locker compartments are intelligently assigned to users. As Figure 3 shown, the intelligent assignment includes fixed locker assignment and shared locker assignment. The fixed locker assignment method is based on the priority queue and the elastic recycling mechanism. The priority queue is sorted according to the user's historical usage rate (average daily usage duration × frequency), and users with high priority are assigned first. At the same time, to prevent users from occupying lockers for a long time, an evaluation of usage efficiency is introduced. If the average usage rate of a user in the past 3 days is lower than the threshold, it will be automatically adjusted to a shared locker and the user will be notified. The shared locker assignment depends on the multi-objective optimization model, that is, maximizing user satisfaction and minimizing the distance of the item retrieval path. The maximization of user satisfaction is based on the preference label weights, that is, the first label obtained is 1 point, the second is 0.5 points, the third is 0.25 points, and so on, halving in turn. Finally, the sum of the preference label scores is obtained. The larger the value, the higher the user satisfaction of the locker assignment; the minimization of the item retrieval path distance is the distance from the work station to the target locker, which is calculated based on the indoor navigation path. The smaller the distance, the higher the recommended value. In addition, the shared locker assignment is restricted by the cabinet type matching degree and the reservation time limit occupancy of the cabinet (such as 30 minutes / time).

[0061] (1) For the fixed locker assignment

[0062] ① Build a priority queue. Collect the historical usage data of all users, calculate the historical usage rate of each user according to the formula "average daily usage duration × frequency"; and sort all users from high to low according to the historical usage rate to build a priority queue.

[0063] ② Assign fixed lockers. Traverse the priority queue. For each user with high priority, check if there is an available fixed locker. If there is an available fixed locker, assign the fixed locker to the user and record the assignment information.

[0064] ③ Usage efficiency evaluation and elastic recycling. Regularly (for example: every day) check the usage situation of users who have been assigned fixed lockers. Calculate the average usage rate of each user in the past 3 days. If it is lower than the preset threshold (for example, 30%, this threshold can be adjusted according to actual business needs), then automatically adjust the fixed locker of this user to a shared locker. At the same time, the system sends a notification to this user to inform that the type of the locker has been changed to a shared locker.

[0065] In a specific embodiment, assume there are users A, B, and C, and their historical usage rates are calculated as follows:

[0066] User A: The average daily usage duration is 2 hours, and the frequency is 5 times / week. The historical usage rate = 2 × 5 = 10;

[0067] User B: The average daily usage duration is 1 hour, and the frequency is 3 times / week. The historical usage rate = 1 × 3 = 3;

[0068] User C: The average daily usage duration is 3 hours, and the frequency is 4 times a week. The historical usage rate = 3×4 = 12;

[0069] Construct a priority queue in the order of C, A, B. If there is an available fixed cabinet at this time, it will be allocated to the user C with the highest priority first. Check the user usage situation after one week. Assume that the average usage rate of user A in the past 3 days is 25%, which is lower than the preset threshold of 30%. Then adjust the fixed cabinet of user A to a shared cabinet and notify user A.

[0070] (2) Regarding the allocation of shared cabinets

[0071] ① Data collection. Collect the preference label data of all users. Each user may have multiple preference labels, such as "close to the reserved workbench", "close to the elevator", "frequently on business trips", etc. Assign corresponding scores to each preference label. The first label is 1 point, the second is 0.5 points, the third is 0.25 points, and so on, halving each time. Determine the workbench location of each user and the locations of all lockers for calculating the distance from the workbench to the target locker.

[0072] ② Calculate the user satisfaction score. For each user, calculate the sum of the preference label scores according to their preference labels. For example, user A has three preference labels, and the sum of their preference label scores is 1 + 0.5 + 0.25 = 1.75 points.

[0073] ③ Calculate the distance of the pick-up path. Based on the indoor navigation path algorithm, calculate the distance from each user's workbench to each shared cabinet.

[0074] ④ Solve the multi-objective optimization model. Construct a multi-objective optimization model with the objective function of maximizing the user satisfaction score and minimizing the pick-up path distance. For each shared cabinet, calculate the comprehensive score of each user.

[0075] ⑤ Allocate shared cabinets. Traverse all shared cabinets. For each shared cabinet, allocate it to the user with the highest comprehensive score. At the same time, consider the constraint of the cabinet type matching degree to ensure that the user's needs assigned match the cabinet type. For example, some shared cabinets may be specifically used for storing large items and can only be allocated to users with such needs. In addition, consider the constraint of the limited reservation time of the cabinet. The reservation time limit for each user to reserve a shared cabinet is 30 minutes, and ensure that the current user's reservation time does not conflict with other users' reservation times when allocating.

[0076] In a specific embodiment, assume there are user D, user E, and user F, and three shared cabinets S1, S2, and S3 respectively.

[0077] The preference labels of user D are "close to the reserved workbench" and "close to the elevator", and the sum of the preference label scores is 1 + 0.5 = 1.5 points.

[0078] The preference label of User E is "Frequently on business trips", and the sum of the preference label scores is 1 point.

[0079] The preference labels of User F are "Close to the reserved work station", "Close to the pantry", and "Quiet area", and the sum of the preference label scores is 1 + 0.5 + 0.25 = 1.75 points.

[0080] Assume that through the calculation of the indoor navigation path:

[0081] The distance from User D to S1 is 10 meters, to S2 is 15 meters, and to S3 is 12 meters.

[0082] The distance from User E to S1 is 18 meters, to S2 is 14 meters, and to S3 is 16 meters.

[0083] The distance from User F to S1 is 13 meters, to S2 is 11 meters, and to S3 is 17 meters.

[0084] Let the weight of the satisfaction score be 0.6 and the weight of the distance be 0.4.

[0085] Calculate the comprehensive score of User D for S1: 0.6×1.5 + 0.4×(1 / 10) = 0.9 + 0.04 = 0.94.

[0086] Calculate the comprehensive score of User D for S2: 0.6×1.5 + 0.4×(1 / 15) ≈ 0.9 + 0.027 = 0.927.

[0087] Calculate the comprehensive score of User D for S3: 0.6×1.5 + 0.4×(1 / 12) ≈ 0.9 + 0.033 = 0.933.

[0088] Similarly, calculate the comprehensive scores of User E and User F for each shared cabinet. Finally, shared cabinet S1 is assigned to User D (assuming its comprehensive score is the highest), S2 is assigned to User F (assuming its comprehensive score is the highest among the candidate users for S2), and S3 is assigned to User E (assuming its comprehensive score is the highest among the candidate users for S3). At the same time, check the cabinet type matching degree and the cabinet reservation time limit occupancy constraint to ensure the rationality of the assignment.

[0089] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for allocating smart lockers in a shared office scenario, characterized in that: The following steps are involved: Step 1: Collect data and preprocess the data. Step 2: Feature modeling; Step 3: Dynamic threshold adjustment and resource optimization; Step 4: Intelligent allocation of lockers.

2. According to the method for allocating smart lockers in a shared office scenario according to claim 1, it is characterized in that: In step 1, the collected data includes user portrait, environmental data, and cabinet status information.

3. According to the method for allocating smart lockers in a shared office scenario as described in claim 1, it is characterized in that: In step 1, data preprocessing includes data cleaning, data standardization and data encoding.

4. According to the method for allocating smart lockers in a shared office scenario as described in claim 1, it is characterized in that: In step 2, the feature modeling is to construct an association matrix, which is as follows: Step 2.1, initialize the association matrix: if there are m users and n lockers, then the association matrix A is an m×n matrix, and all elements in the matrix are 0 at the beginning; Step 2.2: Fill in the association matrix: traverse the usage records of each user, and for the locker j used by user i, update the element A in the association matrix ij ; If user i has only used locker j once, then A ij =1; if user i has used locker j multiple times, then A ij The value of is the number of times used; Step 2.3: Normalize the association matrix: divide each element by the total number of times the user has used the locker.

5. According to the method for allocating smart lockers in a shared office scenario as described in claim 1, it is characterized in that: In step 3, the threshold judgment time range is T1, and the server automatically calculates the average usage rate of T1, average usage rate = total usage time / total available time × 100%; the average usage rate of each locker is compared with the upper and lower thresholds preset during the peak period. If it is lower than the lower threshold, the locker type is converted, and if it is higher than the upper threshold, the locker type remains unchanged; if the same locker grid contains 1 / 2 of the lockers with a usage rate higher than the average usage rate for three consecutive times during the peak period, the system triggers a message reminder to the management personnel, and the same type of lockers are added to the same area.

6. The method for allocating smart lockers in a shared office scenario according to claim 5, wherein: If the average usage rate is within the peak threshold range, the off-peak period is judged. If the off-peak period is lower than the lower limit, the locker type is adjusted, that is, it is fixedly converted to shared. Vice versa, if the off-peak period is higher than the upper limit within the threshold range, the locker type remains unchanged; if the locker usage rate is less than 5% for N consecutive days and there is no weight change, it is marked as a "zombie locker" to trigger recycling or reallocation.

7. The method for allocating smart lockers in a shared office scenario according to claim 1, wherein: In step 4, the intelligent allocation includes fixed cabinet allocation and shared cabinet allocation. The fixed cabinet allocation method is based on priority queue and elastic recovery mechanism. The priority queue is sorted by the user's historical usage rate and introduces the evaluation of usage efficiency. If the user's usage rate is lower than the threshold within the specified time, it is automatically adjusted to a shared cabinet and the user is notified; the shared cabinet allocation is based on a multi-objective optimization model and is subject to the cabinet type matching degree and the cabinet reservation time limit occupancy constraints.

8. The method for allocating smart lockers in a shared office scenario according to claim 7, wherein: The specific process of fixed cabinet allocation is as follows: ① Build a priority queue: collect historical usage data of all users, calculate the historical usage rate of each user according to the formula "average daily usage time × frequency", sort all users from high to low according to historical usage rate, and build a priority queue; ② Allocate a fixed cabinet: traverse the priority queue, and for each high-priority user, check whether there is an available fixed cabinet. If there is an available fixed cabinet, allocate the fixed cabinet to the user and record the corresponding allocation information; ③ Usage efficiency evaluation and flexible recycling: Regularly check the usage of users who have been assigned fixed lockers, calculate the average usage rate of each user in the past three days, and if it is lower than the preset threshold, the user's fixed locker will be automatically adjusted to a shared locker. At the same time, the system will send a notification to the user to inform him that the locker type has been changed to a shared locker.

9. A method for allocating smart lockers in a shared office scenario according to claim 7 or 8, characterized in that: The specific process of sharing cabinet allocation is as follows: ① Data collection: Collect the preference tag data of all users, assign a corresponding score to each preference tag, determine the location of each user's workstation and locker, and calculate the distance from the workstation to the target locker; ② Calculate user satisfaction scores: For each user, calculate the sum of preference label scores based on their preference labels; ③ Calculate the path distance for picking up items: Based on the indoor navigation path algorithm, calculate the distance from each user's workstation to each shared cabinet; ④Solving the multi-objective optimization model: Construct a multi-objective optimization model with the objective function of maximizing the user satisfaction score and minimizing the path distance for picking up items; ⑤ Allocate shared cabinets: Traverse all shared cabinets, and for each shared cabinet, assign it to the user with the highest comprehensive score.

Images