System and method for service site optimization

Abstract

A system and method for service place optimization. The method for service place optimization can include methods for deriving service place utilization data from video, predicting operational characteristics of a service place, optimizing operational characteristics of a service place, and presenting a user-directed optimization of operational characteristics of a service place.

Description

[0001] Cross-reference to related applications

[0002] This application claims priority to Indian Patent Application 202211036805, filed on June 27, 2022, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The various embodiments disclosed herein generally relate to the design and layout of retail service locations, and more specifically to the optimization of the configuration of retail service locations. Background Technology

[0004] Grocery stores and other retail service locations currently use a mix of manned (“full-service” checkout) and unmanned (“self-service” checkout) point-of-sale (POS) terminals. Manned POS terminals can include traditional manned POS terminals, quick manned POS terminals, or reversible manned POS terminals that can be converted to other types of POS terminals, such as quick or self-service. The current POS equipment configuration in many retail service locations may be inefficient in serving customers waiting to check out. That is, the number and combination of POS terminals at a service location may not be well-matched to customer arrival rates and the number of items in customers' shopping baskets. This inefficiency may be due to the number of manned and unmanned POS terminals present at the location and the number of POS terminals operating at a given time during a workday. Furthermore, the effective number and combination of POS terminals at a service location may vary depending on the time of day, the day of the week, and other factors such as holidays, special events, and retailer or manufacturer promotions. The result of this inefficient allocation of POS terminals is reduced customer throughput, slower speed, excessive capital expenditure on unnecessary POS terminals, and an overabundance of goods on largely unused POS terminals. These inefficiencies can lead to customer dissatisfaction, increased operating costs for businesses, decreased sales, and an overall reduction in profit margins.

[0005] This disclosure aims to overcome one or more of the challenges described above. Summary of the Invention

[0006] According to certain aspects of this disclosure, systems and methods for optimizing service locations are disclosed.

[0007] In one embodiment, a computer-implemented method is disclosed for deriving service location utilization data from video in a method for optimizing service locations, the method comprising: receiving one or more video data streams from one or more imaging devices located at a retail service location; uniquely identifying one or more customers of the retail service location within the one or more video data streams; determining one or more metrics of the utilization of the retail service location for each uniquely identified customer of the retail service location; and aggregating each of the one or more metrics determined for each uniquely identified customer of the retail service location into one or more aggregate metrics of the utilization of the retail service location.

[0008] In one embodiment, a computer-implemented method is disclosed for predicting operational characteristics of a service location in a method for service location optimization. The method includes: receiving settings for one or more specified operational characteristics of the service location; calculating latency associated with each of at least one manned service channel and at least one unattended service channel using video data collected from the service location; deriving derived operational characteristics of the service location based on the one or more specified operational characteristics; and reporting another derived operational characteristic of the service location to a user.

[0009] In one embodiment, a computer-implemented method is disclosed for optimizing the operational characteristics of a service location in a method for service location optimization. The method includes: receiving data on one or more operational characteristics of one or more reference service locations; receiving target values ​​for additional operational characteristics of a target service location from a user; receiving data from one or more operating point-of-sale (POS) terminals of the target service location and / or receiving data from the reference service locations; calculating a predicted service rate based on the received operational characteristics and the received data from the operating POS terminals; calculating an average value of the additional operational characteristics of each of the operating POS terminals; selecting a POS terminal whose calculated average value of the additional operational characteristics matches the target value for the additional operational characteristics as a representative POS terminal; and determining the operational characteristics of the target POS terminal of the target service location as the operational characteristics of the representative POS terminal.

[0010] In one embodiment, a computer-implemented method is disclosed for presenting a user interface for user-oriented optimization of operational characteristics of a service location in a method for service location optimization, the method comprising: displaying a user selection user interface, thereby allowing the user to select operational characteristics of the service location to be optimized, and allowing the user to specify one or more parameters for controlling the optimization based on: deriving the derived operational characteristics of the service location based on the one or more specified operational parameters regarding each of at least one manned service channel and at least one unattended service channel using video data collected from the service location; and displaying a results user interface, the results user interface displaying the results of the optimization, the results user interface including a results summary and results details.

[0011] Additional objectives and advantages of the disclosed embodiments will be set forth in part in the description which follows, and in part will be apparent from the description, or may be learned by practice of the disclosed embodiments. The objectives and advantages of the disclosed embodiments will be realized and obtained by means of the elements and combinations particularly pointed out in the appended claims.

[0012] It should be understood that the foregoing general description and the following detailed description are exemplary and illustrative only, and do not limit the claimed disclosed embodiments. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate various exemplary embodiments and, together with the specification, serve to explain the principles of the disclosed embodiments.

[0014] Figure 1 An exemplary overview of retail service locations in a method for optimizing service locations, according to one or more implementation schemes, is depicted.

[0015] Figure 2 An exemplary system infrastructure for service site optimization according to one or more implementation schemes is described.

[0016] Figure 3 The process flow of a method for optimizing service locations according to one or more implementation schemes is described.

[0017] Figure 4 Clustering mappings of retail service locations are depicted in a method for optimizing service locations according to one or more implementation schemes.

[0018] Figure 5 A flowchart is depicted to model customer flow redistribution in a service location optimization approach based on one or more implementation schemes.

[0019] Figure 6A flowchart is depicted in a method for optimizing service locations according to one or more implementation schemes to calculate optimized waiting times.

[0020] Figure 7 A flowchart is depicted for calculating the optimized average number of customers waiting in a queue in a service location optimization method according to one or more implementation schemes.

[0021] Figure 8 A flowchart is depicted for calculating the optimized average utilization rate of POS terminals in a service location optimization method according to one or more implementation schemes.

[0022] Figure 9 A flowchart is depicted for calculating optimized labor costs in a service location optimization method according to one or more implementation schemes.

[0023] Figure 10 A graph depicts the calculated optimized labor costs in a method for optimizing service locations according to one or more implementation schemes.

[0024] Figure 11 The process flow of a method for optimizing service locations according to one or more implementation schemes is described.

[0025] Figure 12 A flowchart is provided illustrating a method for optimizing service locations based on one or more implementation schemes.

[0026] Figure 13 A simplified functional block diagram of a computer, which is a device configured to perform a service location optimization method according to one or more embodiments, is shown.

[0027] Figure 14 An exemplary user interface for service location optimization according to one or more implementation schemes is depicted.

[0028] Figure 15 An exemplary user interface for service location optimization according to one or more implementation schemes is depicted. Detailed Implementation

[0029] The various implementation schemes disclosed herein generally involve optimizing the configuration of retail service locations.

[0030] The terms used below may be interpreted in their broadest and most reasonable manner, even when used in conjunction with a detailed description of certain specific examples of this disclosure. In fact, some terms may even be emphasized below; however, any term intended to be interpreted in any limiting manner will be disclosed and specifically defined in this Detailed Description section.

[0031] As discussed above, the number and combination of POS terminals at a service location may not be well-matched with customer arrival rates and the number of items in customers' shopping baskets. One or more implementations can address this inefficiency by optimizing the number, combination, and operating schedule of POS terminals at the service location. This can be achieved by measuring and monitoring operational characteristics of the service location, including, for example, customer arrival rates, items selected per customer, customer wait times, and service times at the POS terminals. These characteristics can be measured by analyzing video data, data from other sensors, and data from the POS terminals themselves. Service location operators can then obtain performance metrics for the service location, models and predictions of key characteristics of the service location based on user-selected criteria, and suggested optimizations for POS terminal configuration and operation based on user-selected criteria.

[0032] Any suitable system infrastructure can be placed in place to optimize the configuration of retail service locations. Figure 1 , Figure 2 and Figure 13 The following discussion provides a brief, general description of a suitable computing environment in which this disclosure may be implemented. In one embodiment, any of the disclosed systems, methods, and / or graphical user interfaces can be implemented by... Figure 1 , Figure 2 and Figure 13 The computing systems described herein are consistent with or similar to computing systems that perform or implement the functions. Although not required, aspects of this disclosure are described in the context of computer-executable instructions, such as routines executed by data processing devices (e.g., server computers, wireless devices, and / or personal computers). Those skilled in the art will understand that aspects of this disclosure can be practiced with other communication, data processing, or computer system configurations, including: internet devices, handheld devices (including personal digital assistants (“PDAs”), wearable computers, various cellular or mobile phones (including Voice over IP (“VoIP”) phones), dumb terminals, media players, gaming devices, virtual reality devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, set-top boxes, network PCs, microcomputers, mainframe computers, etc. In practice, the terms “computer,” “server,” etc., are generally used interchangeably herein and refer to any of the aforementioned devices and systems, as well as any data processor.

[0033] Various aspects of this disclosure can be embodied in a dedicated computer and / or data processor specifically programmed, configured, and / or constructed to perform one or more of the computer-executable instructions described in detail herein. While aspects of this disclosure (such as certain functions) are described as exclusively implemented on a single device, this disclosure can also be practiced in a distributed environment where functions or modules are shared between entirely different processing devices linked by a communication network such as a local area network (“LAN”), a wide area network (“WAN”), and / or the Internet. Similarly, the techniques described herein involving multiple devices can be implemented in a single device. In a distributed computing environment, program modules can reside in both local and / or remote memory storage devices.

[0034] The aspects of this disclosure can be stored and / or distributed on non-transitory computer-readable media, including magnetic or optically readable computer disks, hardwired or pre-programmed chips (e.g., EEPROM semiconductor chips), nanotechnology memories, biological memories, or other data storage media. Alternatively, computer-implemented instructions, data structures, screen displays, and other data according to the aspects of this disclosure can be distributed over a period of time via the Internet and / or other networks (including wireless networks) on propagation signals on a propagation medium (e.g., electromagnetic waves, sound waves, etc.), and / or the computer-implemented instructions, data structures, screen displays, and other data can be made available on any analog or digital network (packet switching, circuit switching, or another scheme).

[0035] As used herein, "machine learning model" generally encompasses instructions, data, and / or models configured to receive input and apply one or more of weights, biases, classifications, or analyses to that input to generate output. Output may include, for example, a classification of the input, an analysis based on the input, a design, processing, prediction, or recommendation associated with the input, or any other suitable type of output. Machine learning models are typically trained using training data (e.g., empirical data and / or samples of input data) that is fed into the model to establish, adjust, or modify one or more aspects of the model, such as weights, biases, criteria used to form classifications or clusters, etc. Aspects of a machine learning model may operate on the input linearly and in parallel via a network (e.g., a neural network) or via any suitable configuration.

[0036] The execution of a machine learning model can include deploying one or more machine learning techniques, such as linear regression, logistic regression, random forests, gradient boosting machines (GBM), deep learning, and / or deep neural networks. Supervised and / or unsupervised training can be employed. For example, supervised learning can include providing training data and corresponding labels (e.g., ground truth) as the training data. Unsupervised methods can include clustering, classification, etc. K-means clustering or K-nearest neighbors can also be used, and these can be supervised or unsupervised. A combination of K-nearest neighbors and unsupervised clustering techniques can also be used. Any suitable type of training can be used, such as stochastic, gradient boosting, random seeding, recursive, epoch-based, or batch-based training, etc.

[0037] Retail service venues

[0038] Figure 1 An exemplary overview of a retail service location 100 according to one or more embodiments is depicted, and techniques for optimizing the service location are disclosed for this retail service location. Figure 1 As shown, retail service location 100 may include multiple manned or "full-service" point-of-sale (POS) terminals 110, such as Figure 1 The manned POS terminals 110a-110c shown are illustrated. Retail service locations 100 may also include multiple unattended or "self-service" POS terminals 150, such as... Figure 1The unattended POS terminals 150a-150b are shown. Each attended POS terminal 110 can be a traditional attended or "full-service" POS terminal, a quick attended POS terminal, or a reversible attended POS terminal that can be converted to other types of POS terminals (such as quick or self-service). The unattended POS terminals 150 can be provided individually (each unattended POS terminal has a separate entrance and queue) or together in a "bullpen" such as having a single entrance and queue. Each of the POS terminals 110 and 150 can be operational (i.e., "on") or not operational (i.e., "off"). Similarly, the attended POS terminals 110 can have a separate queue for each POS terminal 110 or a single queue for multiple POS terminals 110. Each of the POS terminals 110 and 150 can include tags 145 presenting information about the POS terminals 110 and 150, such as channel identifiers and / or operational status of the POS terminals 110 and 150. The operating states of POS terminals 110 and 150 can be, for example, "on," "off," "ready for the next customer," "customer needs assistance," or "requires equipment service." Each of POS terminals 110 and 150 may also include, for example, a POS system 135 and a fixture 130, such as a shelf, counter, or table, on which customers can display the items they wish to purchase. The configuration of each POS terminal 110 and 150 (such as including a scanner, conveyor belt, payment keypad, currency receiver and dispenser, display, etc.) can be any suitable configuration as would be understood by those skilled in the art. Furthermore, the "on" operating state of the manned POS terminal 110 can be operated by a sales clerk or cashier 140.

[0039] Each operating POS terminal 110 and 150 may have a line of customers 115 waiting to complete their purchase. Each customer 115 may be carrying multiple items 125 to purchase, such as items that may be held in their hands and / or placed in a shopping basket or cart 120.

[0040] Activities within the retail service location 100 can be detected and measured by various sensors located within the retail service location 100, such as the arrival of customer 115, customer 115 waiting in queues at each POS terminal 110 or 150, customer wait time at each POS terminal 110 or 150, average service time at each POS terminal 110 or 150, and the quantity or combination of items 125 to be purchased by each customer 115. For example, the retail service location 100 may be equipped with one or more cameras 155, one or more motion detectors 160, and one or more sensors 165 for detecting identification tags attached to items 125, such as via radio frequency identification (RFID). For example, the presence and movement of customer 115 can be determined using cameras 155 and / or motion detectors 160. When a customer 115 moves around the retail service area 100 or during the checkout process when items 125 are processed at the POS system 135, the quantity or combination of items 125 in the customer's shopping basket or cart 120 can be determined, for example, by optical recognition via camera 155 and / or by remotely reading the identification tags attached to items 125 by sensor 165. Alternatively, real-time object recognition in images captured by camera 155 can identify items 125 selected and placed in the shopping basket or cart 120 by the customer 115.

[0041] Determining the operational characteristics of retail location 100 may include, for example: receiving one or more video data streams from one or more imaging devices located at the retail location; uniquely identifying one or more customers at the retail location within the one or more video data streams; determining one or more metrics of the utilization of the retail location for each uniquely identified customer at the retail location; and aggregating each of the one or more metrics determined for each uniquely identified customer at the retail location into one or more aggregate metrics of the utilization of the retail location.

[0042] Retail service location data can be aggregated at any desired frequency (hourly, daily, weekly, etc.) within any desired current or past time period (current day, week, year, etc., the most recent day, week, year, etc., and a selected day, week, year, etc. in the past). This retail service location data includes, for example, the number of POS terminals 110 or 150 by type (attended or unattended), traffic to retail service locations 100, and sales volume of retail service locations 100.

[0043] Modeling of retail service venue operations

[0044] The owner or operator of retail location 100 may wish to better understand the operation of retail location 100 in order to optimize the area of ​​POS terminals 110 and 150, i.e. the “front end” of retail location 100, for example, through more efficient scheduling or operation of existing POS terminals 110 and 150 or through improved design of the front end of retail location 100, and to model the return on investment (ROI) of retail location 100 relative to labor costs, operating hours, floor space utilization, etc.

[0045] This improved understanding of the operation of retail location 100 can be obtained by applying modern statistical modeling methods, such as queuing theory modeling. For example, the owner or operator of retail location 100 can better understand the service level of retail location 100 based on measured, calculated, modeled, or predicted average wait times and average service times. The performance of retail location 100 can be measured on any suitable time scale, such as minutes, hours, days, or time periods within a day, and can include peak and off-peak times of day, weekends and weekdays, public holidays and regular days, specific advertising and promotional days, etc. This analysis can not only indicate the potential performance gains of a particular retail location 100, but also proactively identify retail locations 100 that can benefit from such analysis.

[0046] The data used as input to these statistical methods may include, for example, information categorized by the type of POS terminal 110 and 150 (e.g., unattended or “self-service” POS terminal, traditional attended or “full-service” POS terminal, quick attended POS terminal, or reversible attended POS terminal that can be converted to other types of POS terminals, such as quick self-service). Data may be collected within any appropriate time frame (e.g., throughout the year (365 days)) and may cover a full day or any part of the operating hours of retail service location 100 (e.g., 6:00 a.m. to 11:59 p.m.). The collected data may include, for example, customer arrivals per hour, number of transactions per hour, number of items per transaction, total amount spent per shopping basket (“basket size”), hourly traffic in the store, etc.

[0047] The sample data of 100 retail service locations are shown in Table 1 below.

[0048] Table 1

[0049]

[0050] Depending on the type of POS terminal, data collected at retail service locations 100 can provide insights into customer wait times and average service times at locations such as POS terminals 110 and 150. For example, as shown in Table 2, a traditional manned POS terminal 110 may have a longer wait time but a shorter service time compared to an unattended POS terminal 150.

[0051] Table 2

[0052] Time (in seconds) Waiting time Service Hours Total Time Manned - Fast 3.4 2.3 5.7 Manned - Traditional 2.7 2.1 4.8 Reversible 2.8 2.2 5.0 self service 0.1 2.7 2.8

[0053] Statistical and modeling techniques (e.g., queuing theory modeling) can be used to further explore insights such as these, as discussed in more detail below.

[0054] Queue theory modeling can be used to understand and design the workflows of multiple retail service locations 100. This may involve using mathematical methods to capture, define, and optimize relationships such as "service time" and "items in shopping baskets." These relationships can be used to generate and display desired metrics, such as "service rate capacity," and recommended store / location configurations, such as the optimized number and combination of manned and unmanned service lanes. Such analysis can yield a better understanding of the operation of retail service locations 100 and can be used to recommend improved configurations and operations for retail service locations 100.

[0055] This application of queuing theory modeling can include calculating the "service rate capacity" (SRC) at each type of POS terminal 110 or 150 by capturing the relationship between "service time" and the number of items in customer 115's shopping basket 120. Queueing theory modeling can take one or more inputs regarding the number of operating POS terminals 110 or 150, the arrival rate (i.e., the number of customers 115 joining the queue at POS terminal 110 or 150), and the SRC. Queueing theory modeling can produce the following as outputs: the number of customers 115 waiting in the queue, the waiting time of customers 115 waiting in the queue, the probability of a specific number of customers 115 waiting in the queue, the probability that the waiting time of customers 115 exceeds a specific threshold (30 seconds, 1 minute, 2 minutes, 5 minutes, etc.), and so on.

[0056] Generalized queuing theory modeling might be based on formulas such as:

[0057] λ is the average arrival rate of customers 115 at retail service location 100. That is, the average number of customers 115 arriving at retail service location 100 per unit of time.

[0058] μ is the average service rate. That is, the average number of customers that can be served per unit of time.

[0059] The average utilization rate of systems such as POS terminals 110 and 150 is ρ = λ / μ.

[0060] The average number of customers 115 in the system of retail service location 100 is L = λ / (μ-λ).

[0061] The average number of customers waiting in the queue (that is, in queues 110 and 150 at POS terminals) is L. Q =ρL.

[0062] The average time customer 115 spends at retail service location 100 (including service time at POS terminals 110 and 150) is W = 1 / (μ – λ).

[0063] Customer 115 spends an average of W waiting time in queues (that is, in queues 110 and 150 at POS terminals). Q =ρW.

[0064] At a given time, the probability that n customers 115 are located in retail service location 100 is P. n =(1-ρ)ρ n .

[0065] Furthermore, these formulas can be derived from, for example, by Figure 2 The optimization module 235 described is used for computational optimization of retail service locations 100:

[0066] The average number of customers waiting in the queue (N) q )=(ρ / (1-ρ)) / (1+Q), where ρ=r / (percentage of POS terminals 110 and 150 in operation) and r=(λ) / (service rate capacity).

[0067] Average waiting time in the queue (T) q )=(N q / λ)*60.

[0068] Average utilization rate of POS terminals 110 and 150 = ((λ) / ((service rate capacity)*(percentage of POS terminals 110 and 150 in operation)))*100.

[0069] The number of POS terminals 110 and 150 in operation = the total number of POS terminals 110 and 150 * the percentage of POS terminals 110 and 150 in operation.

[0070] This analysis can yield insights such as the additional interaction between varying customer arrival rates and different transaction sizes, the predictability of this variation by time of day and day of the week, and how it affects the wait times of customers queuing for 115, as well as the predictable variations in transaction processing speed between attended POS terminals 110 and unattended POS terminals 150. For example, as discussed above with respect to Table 2, attended POS terminals 110 may consistently process transactions of the same size faster than unattended POS terminals 150. These insights can allow for regulated utilization and wait times at various POS terminals 110 or 150, determining which type of POS terminal 110 or 150 might be best suited for optimal utilization, and planning service levels for POS terminals 110 or 150 to cope with peak and off-peak hours of the day, weekends and weekdays, public holidays and regular days, etc.

[0071] Furthermore, this analysis can provide insights into the relationship between the quantity of items 125 in a customer's shopping basket or cart 120 and the customer's choice between an attended POS terminal 110 and an unattended POS terminal 150. For example, a customer 115 with fewer items 125 in their shopping basket or cart 120 may be more likely to choose an unattended POS terminal 150, while a customer with more items in their shopping basket or cart 120 may be more likely to choose an attended POS terminal. Additionally, this analysis can provide insights into how the total number of transactions per hour in the retail service location 100 and the varying quantities of items 125 in a customer 115's shopping basket or cart 120 might affect the accuracy of a customer 115's choice between an attended POS terminal 110 and an unattended POS terminal 150 when their basket coefficient is at the margin.

[0072] Insights into the arrival rate of customer 115 and different transaction sizes, the waiting time of customer 115 in the queue, and the choice of customer 115 between an attended POS terminal 110 and an unattended POS terminal 150 can allow for the simulation of the probability of the number of waiting customers 115 and the total waiting time in the queue.

[0073] Based on these insights and simulations, one or more predictions can be made regarding the operation of retail location 100. For example, the number and type of manned or unattended POS terminals 110 or 150 required to meet customer service requirements can be predicted. Such predictions can allow the operator of retail location 100 to design the checkout experience for customer 115 around customer wait times, thereby making it possible to manage service hours at retail location 100 and more efficiently allocate capital expenditures on manned and unattended POS terminals 110 and 150. Additionally, such predictions can allow the operator of retail location 100 to plan the working hours of staff or cashiers 140 based on time of day and day of the week to successfully meet customer service requirements.

[0074] Predicting the operational characteristics of a retail service location 100 may include, for example, receiving settings for one or more specified operational characteristics of the retail service location 100, deriving another operational characteristic of the service location based on one or more specified operational characteristics, and reporting the derived operational characteristic of the service location to a user. Specified operational characteristics may be received from a user via a user interface or from a database (such as database 200) containing past operational characteristics of the retail service location 100 and / or additional data on the retail service location 100. The operational characteristics of the retail service location 100 may include one or more of the following: multiple manned point-of-sale (POS) terminals present at the service location, multiple unattended POS terminals present at the service location, operating schedules of manned and unattended POS terminals, customer arrival rate at the service location, and the average number of service items selected by customers at the service location.

[0075] These technologies enable more efficient design of transaction areas (the areas where purchase transactions are completed in retail service locations 100) to achieve better customer service, more efficient capital allocation for purchasing POS equipment, reduced inventory days, and reduced profit loss due to aging products located in largely unused POS terminals.

[0076] Figures 2 to 13 The following discussion describes the technical environment and related methods that enable this application of queuing theory modeling.

[0077] Retail Service Venue Optimizer

[0078] Figure 2 An exemplary system infrastructure for service location optimization according to one or more implementation schemes is described. For example... Figure 2 As shown, the system for optimizing service locations can be used in, for example... Figure 1 Of the 100 retail service locations described. Specifically, such as Figure 1 As shown, a retail service location 100 may include one or more manned POS terminals 110 (also referred to as “full-service” checkouts), one or more unattended POS terminals 150 (also referred to as “self-service” checkouts), and various sensors located within the retail service location 100, such as one or more cameras 155, one or more motion detectors 160, one or more radio frequency identification (RFID) sensors 165, etc. Data from POS terminals 110 and 150 and from the various sensors can be provided to a service location optimizer 260. It should be understood that although the service location optimizer 260 is shown communicating with cameras 155, sensors 160, 165, POS terminals 110, 150, the service location optimizer 260 can perform on-site operations at the service location or receive data from those devices via any network such as the Internet. In either case, the service location optimizer 260 can be configured to receive input data locally or via a network for various measurements, modeling, prediction, and optimizations of the operation of the retail service location 100 as discussed above. These functions will be described below regarding... Figures 3 to 12 To describe in more detail.

[0079] The service location optimizer 260 can generate metrics, predictions, models, and recommendations for the retail service location 100, and may include, for example, a service location metrics calculator 210, a database 220, a predictive modeling module 230, an optimization module 235, a reporting module 240, and a user interface module 250.

[0080] The service location metrics calculator 210 can use information received from POS terminals 110 and 150 and from various sensors to calculate various metrics for the retail service location 100. These metrics include, for example, the number of customers 115 arriving per time period (minute, hour, time of day, etc.), the total number of transactions per time period (minute, hour, time of day, day, etc., aggregated for all retail service locations 100, each POS terminal 110 or 150, each type of POS terminal 110 or 150, etc.), the average number of items per transaction (aggregated per minute, hour, time of day, day, etc., aggregated for all retail service locations 100, each POS terminal 110 or 150, each type of POS terminal 110 or 150, etc.), and the total amount spent per customer 115 (e.g., per minute, hour, time of day, day, etc., for all retail service locations). Average basket factor (aggregated by 100, 110 or 150 per POS terminal, 110 or 150 per type of POS terminal, etc.), customer traffic in retail service locations 100 per time period (minute, hour, time of day, day, etc.), average number and percentage of operating POS terminals 110 and 150 per time period (per minute, hour, time of day, day, etc., aggregated for all retail service locations 100, 110 or 150 per POS terminal, 110 or 150 per type of POS terminal, etc.), distribution of all transactions by POS terminal type (per minute, hour, time of day, day, etc., 110 or 150 per type of POS terminal), percentage of total spending by all customers 115 by POS terminal type (per minute, hour, time of day, day, etc., 110 or 150 per type of POS terminal, etc.).

[0081] Additionally, the service area measurement calculator 210 can use information received from various sensors to measure the movement of customer 115 within the retail service area 100, including, for example, entering queues at POS terminals 110 or 150, switching between queues at POS terminals 110 or 150, leaving queues at POS terminals 110 or 150, the time spent in queues at POS terminals 110 or 150, leaving the retail service area 100 without entering queues at POS terminals 110 or 150, etc. These measurements can be combined with information received from POS terminals 110 and 150 to categorize these movements by, for example, the total number of items 125 in shopping basket 120, or the total amount spent on items 125 in shopping basket 120.

[0082] Information received from POS terminals 110 and 150 and from various sensors, as well as measurements calculated by the service location measurement calculator 210, can be stored in database 220. Database 220 can store data related to a specific retail service location 100 and / or multiple retail service locations 100.

[0083] The predictive modeling module 230 can model using the measurement results and metrics generated by the service location metric calculator 210, and then predict one or more metrics of the retail service location 100 based on past metrics actually measured at the retail service location 100, such as the metrics discussed above, or one or more metrics of other retail service locations in the retail service location 100, such as those that can be stored in the database 220, as well as user-provided settings for one or more operational characteristics of the retail service location 100. User-provided settings for one or more operational characteristics may include, for example, the total number of transactions, the number of manned POS terminals 110 present at the retail service location 100, the number of unattended POS terminals 150 present, and the schedules of POS terminals 110 and 150 operating according to the time of day, the day of the week, and other factors such as holidays, special events, and retailer or manufacturer promotions. User-provided settings may also include customer characteristics, including, for example, the arrival rate of customer 115, the average number of items 125 in each shopping basket 120, the average total amount spent on items 125 in each shopping basket 120, etc. Customer characteristics can be specified based on time of day, day of week, and other factors such as holidays, special events, and retailer or manufacturer promotions. Alternatively, these settings can be automatically derived based on past performance data from retail location 100 or other retail locations within retail location 100, such as data that can be stored in database 220.

[0084] The predictive modeling module 230 can generate predictive metrics by any suitable means, including, for example, queuing theory, machine learning (such as by machine learning module 232), statistical analysis, artificial intelligence, or any suitable combination.

[0085] The optimization module 235 can use the measurement results and metrics generated by the service location metrics calculator 210 to generate recommended operating characteristics for the retail service location 100 to optimize one or more metrics of the retail service location 100, such as those discussed above. Recommended operating characteristics may include, for example, labor costs, the number of manned POS terminals 110 present at the retail service location 100, the number of unattended POS terminals 150 present, and the schedules of POS terminals 110 and 150 operating based on time of day, day of week, and other factors such as holidays, special events, and retailer or manufacturer promotions.

[0086] The optimization module 235 can generate recommended operating characteristics for the retail service location 100 by any suitable means, including, for example, queuing theory, machine learning (such as through machine learning module 237), statistical analysis, artificial intelligence, or any suitable combination.

[0087] The reporting module 240 can generate reports, such as text reports or graphical reports, to present the metrics, predictions, models, and recommended optimizations of the retail service venue 100 generated by the venue optimizer, service venue metrics calculator 210, database 220, prediction module 230, or optimization module 235.

[0088] User interface module 250 can generate and present an interactive graphical user interface (including optimizations of the retail service location 100’s metrics, predictions, models and recommendations generated by the location optimizer, service location metrics calculator 210, prediction module 230 or optimization module 235) to the user, as well as an interactive graphical user interface for receiving user input (e.g., user-provided settings for one or more operating characteristics of the retail service location 100).

[0089] The service location optimizer 260 can connect to other devices or servers in other locations via a computer network 265, which can be any suitable type of computer network, such as a local area network (LAN), wide area network (WAN), wireless network, or the Internet. For example, the service location optimizer 260 can transmit data including metrics, predictions, models, and recommended optimizations of the retail service location 100 generated by the location optimizer, service location metrics calculator 210, prediction module 230, or optimization module 235 to a central office 270 (where planning and control of multiple retail service locations 100 can be performed), or to a back-end 280 specific to a particular retail service location 100, where planning and control of that particular retail service location 100 can be performed. Furthermore, such information can be made available to any network-connected device, such as a mobile device application, via a reporting portal 290.

[0090] Machine Learning

[0091] As discussed in further detail below, the service location optimizer 260 may perform one or more of the following operations: (i) generating, storing, training, or using an optimized machine learning model configured to generate recommendations for the retail service location 100. The service location optimizer 260 may include the machine learning model and / or instructions associated with the machine learning model, such as instructions for generating the machine learning model, training the machine learning model, using the machine learning model, etc. For example, these functions may be provided by the machine learning module 232 of the modeling module 230 or the machine learning module 237 of the optimization module 235, such as... Figure 2 The service location optimizer 260 may include functions for: retrieving data related to the operation of one or more retail service locations 100; analyzing such data to, for example, model and subsequently predict one or more metrics of the retail service location 100, or generating recommended operating characteristics of the retail service location 100 to optimize one or more metrics of the retail service location 100 based on the output of a machine learning model, and / or generating one or more reports or user interfaces to display, for example, the generated model, prediction, or recommendation based on a machine learning model. The service location optimizer 260 may include training data and ground-based or validation data. For example, the service location optimizer 260 may use recorded data from one or more retail service locations 100 (such as data that may be stored in database 220) to train one or more machine learning models. A portion of the data (e.g., 70%) may be used as training data, and the remainder (e.g., 30%) may be used as validation data.

[0092] In some implementations, systems or devices other than the service location optimizer 260 may be used to generate and / or train machine learning models. For example, such a system may include instructions for generating machine learning models, training data, and ground real-time data, and / or instructions for training the machine learning models. The resulting trained machine learning model can then be provided to the service location optimizer 260.

[0093] Generally, machine learning models consist of a set of variables, such as nodes, neurons, and filters, which are adjusted (e.g., weighted or biased) to different values ​​through the application of training data. In supervised learning, for example, where the provided training data is based on known ground reality, training can be performed by feeding training data samples into a model with variables that are, for example, randomly initialized, based on Gaussian noise, a pre-trained model, etc. The output can be compared to the ground reality to determine the error, which can then be backpropagated through the model to adjust the values ​​of the variables.

[0094] Training can be conducted in any suitable manner, such as in batches, and can include any suitable training method, such as stochastic or non-stochastic gradient descent, gradient boosting, random forests, etc. In some implementations, a portion of the training data can be retained during training and / or used to validate the trained machine learning model, for example, by comparing the output of the trained model with ground truth data from that portion of the training data to evaluate the accuracy of the trained model. The training of the machine learning model can be configured such that the machine learning model learns the association between the training data and the ground truth data, such that the trained machine learning model is configured to determine, for example, a model, prediction, or recommendation generated based on the output generated by the machine learning model in response to input recorded data from one or more retail service locations 100 based on the learned associations.

[0095] In some cases, the different training data and / or input data samples may not be independent. For example, a dataset of past performance data of one or more retail service locations 100 (such as that which may be stored in database 220) may include data samples from multiple different time periods for each retail service location 100. Therefore, in some implementations, the machine learning model may be configured to consider and / or determine the relationships between multiple samples.

[0096] For example, in some implementations, the machine learning module 232 of the modeling module 230 or the machine learning module 237 of the optimization module 235 (such as...) Figure 2 The machine learning model described may include a recurrent neural network (“RNN”). Generally, RNNs are a class of feedforward neural networks well-suited for processing input sequences. In some implementations, the machine learning model may include a long short-term memory (“LSTM”) model and / or a sequence-to-sequence (“Seq2Seq”) model. An LSTM model may be configured to generate an output from samples that takes into account at least some previous samples and / or outputs. A Seq2Seq model may be configured to, for example, receive a series of non-optical in vivo images as input and generate a series of locations (e.g., paths) in medical imaging data as output.

[0097] Other aspects of one or more machine learning models and / or how they can be used to optimize the configuration and operation of retail service locations 100 will be discussed in further detail in the methods below. In the methods below, various actions can be described as being performed by... Figure 2The actions are performed or executed by components within the environment (such as the service location optimizer 260 or its components). However, it should be understood that in various embodiments, the various components of the environment 200 discussed above can execute instructions or perform actions including those discussed below. Actions performed by the device can be considered to be performed by processors, actuators, etc., associated with the device. Furthermore, it should be understood that in various embodiments, steps can be added, omitted, and / or rearranged in any suitable manner.

[0098] Retail service venue modeling

[0099] The optimization of the metrics, predictions, models, and recommendations of the retail service venue 100 generated by the service venue optimizer 260 can be part of a cyclical or “closed-loop” feedback process that iterates and optimizes the configuration and operation of the retail service venue 100. Figure 3 The process flow of this method for optimizing service locations according to one or more implementation schemes is described.

[0100] like Figure 3 As shown, the iterative and optimized cyclical process of configuring and operating the retail service location 100 may include, for example, data collection operation 310, metric processing operation 320, modeling operation 330, optimization operation 340, report generation operation 350, reporting operation 360, and control and configuration operation 370. After the control and configuration operation 370 has completed tasks such as applying the optimized configuration and / or operation to the retail service location 100, the process can return to the data collection operation 310 and can be repeated to ensure continued optimal operation of the retail service location 100. In some cases, the control and configuration operation 370 may be accompanied by organic changes in the conditions at the service location, or even replaced by such organic changes (e.g., changes in conditions may trigger the collection of new data for updating the model and optimization, even when the control or configuration has not changed).

[0101] Data collection operation 310 may include, for example, collecting data from various sensors present at retail service locations 100 and 110 and 150. The collected data may be stored, for example, in a database, such as... Figure 2 The database 220 is described.

[0102] The measurement processing operation 320 may include, for example, using information received from POS terminals 110 and 150 and from various sensors, or other data stored in, for example, database 220, to calculate various metrics of the retail service location 100, as discussed above with respect to the service location measurement calculator 210. The generated metrics may be stored, for example, in database 220.

[0103] Modeling operation 330 may include generating various models of the operation of retail service location 100, as discussed above regarding modeling module 230.

[0104] Optimization operation 340 may include, for example, using measurements and metrics of retail service location 100 to generate recommended operating characteristics of retail service location 100 to optimize one or more metrics of retail service location 100, as discussed above with respect to optimization module 235.

[0105] Report generation operations 350 and 360 may include, for example, generating reports, such as text or graphical reports, to present optimizations of metrics, predictions, models, and recommendations for the retail service location 100, as discussed above regarding report module 240, and delivering the generated reports to various business operations and users, such as central office 270, back office 280, or via report portal 290, as discussed above regarding... Figure 2 The discussion.

[0106] Control and configuration operations 370 may include, for example, determining the operating schedule of POS terminals 110 and 150, determining additional POS terminals 110 and 150 to be added to the retail service location 100, determining one or more POS terminals 110 and 150 in the retail service location 100 that will be converted from manned POS terminal 110 to unattended POS terminal 150 or vice versa, determining one or more POS terminals 110 and 150 to be removed from the retail service location 100, and so on.

[0107] After control and configuration operation 370 has completed tasks such as applying optimized configuration and / or operation to retail service location 100, the process can return to data collection operation 310 and can be repeated to ensure continued optimal operation of retail service location 100.

[0108] Prediction and modeling operations

[0109] As discussed above, the process for optimizing the configuration and operation of retail service locations 100 (such as that implemented by the service location optimizer 260) may include generating various metrics, metric predictions, and models regarding the operation of retail service locations 100. These metrics, metric predictions, and models may include, for example, clustering models, arrival rate models, service rate models, traffic redistribution algorithms, algorithms for determining the operation of POS terminals 110 and 150, and one or more optimization algorithms. The various metrics, metric predictions, and models regarding the operation of retail service locations 100 may be derived based on data collected and stored about retail service locations 100, other retail service locations 100, or combinations thereof.

[0110] Clustering models can utilize hourly and store-level metrics to place retail location 100 into one of several clusters similar to retail location 100 based on its configuration and performance. For example, such clustering can be useful when comparing retail location 100 with other similar retail locations 100. For instance, clustering models can be used to group multiple sample locations into distinct categories (e.g., low volume and high volume, etc.). Table 3 below depicts exemplary clusters of sample retail locations.

[0111] Table 3

[0112]

[0113] Figure 4 A user interface display of a clustering map 400 of retail service locations in a service location optimization method according to one or more implementation schemes is depicted. (Example) Figure 4 As shown, each retail location 100 can be represented by graphical elements 430 arranged within a clustering map 400 based on, for example, weekly sales 410 and weekly transaction volume 420. A scale 440 can reflect the clustering of the retail locations 100. Clustering can be performed with respect to the metrics and measured data of the retail locations 100, and can be accomplished through any suitable statistical process (e.g., k-means clustering).

[0114] Clustering models can be derived from Figure 2 The machine learning module 232 described is implemented. As discussed above, a machine learning model can be trained using a dataset, such as one that can be stored in database 220, of past performance data of one or more retail service locations 100. For example, a portion of the past performance data can be used as a training dataset and another portion as a validation dataset to train the model. For example, 70% of the past performance data can be used as a training dataset and the remaining 30% can be used as a validation dataset. A clustering model can be trained using, for example, the average weekly transaction volume and average weekly sales of each retail service location 100. The clustering model can be, for example, an unsupervised machine learning model and can be validated based on silhouette scores.

[0115] Within the user interface display of the clustering map 400, the formatting and rendering of the graphical element 430 (such as size, highlighting, animation, etc.) can indicate information about the associated retail service location 100, such as sales volume, profitability, operational status, etc. Furthermore, user interaction with the graphical element 430 (such as clicking, hovering, etc.) can trigger the display of detailed information about the associated retail service location 100, or can open a dashboard or another user interface, such as the user interface discussed above regarding user interface module 250.

[0116] The arrival rate model can use hourly and store-level metrics to predict customer arrival rates. The modeled customer arrival rate can be used to calculate key performance indicators for retail service locations 100, such as wait times, number of waiting customers, and utilization rates of POS terminals 110 and 150. This model can be performed using any suitable statistical procedure, such as linear regression.

[0117] The arrival rate model can be derived from Figure 2 The machine learning module 232 described is implemented. As discussed above, a machine learning model can be trained using a dataset of past performance data from one or more retail locations 100, such as a dataset that can be stored in database 220. For example, a portion of the past performance data can be used as a training dataset and another portion as a validation dataset to train the model. For example, 70% of the past performance data can be used as a training dataset and the remaining 33% can be used as a validation dataset. The arrival rate can be trained using, for example, the average number of transactions for seasonal periods (standard hours, Christmas and Thanksgiving), weekdays (Monday, Tuesday, etc.), peak periods (peak and off-peak), and the POS type (attended POS terminal 110, unattended POS terminal 150, etc.) for each retail location 100. The arrival rate model can be, for example, a supervised machine learning model and can be validated based on the mean absolute percentage error (MAPE). Furthermore, the root mean square error (RMSE) metric can be used instead of or in combination with MAPE to evaluate the training and validation of the arrival rate model. Customer arrival rate can be correlated with the number of transactions at the retail location 100. Therefore, transaction data can be used to estimate customer arrival rates at any time period at a retail service location.

[0118] The operating levels of POS terminals 110 and 150 can be determined based on collected current or historical data, or can be specified by the user. The operating levels of POS terminals 110 and 150 can be specific to each POS terminal 110 and 150, or can be distributed across POS terminals 110 and 150. However, the total operating level (total or individually) of POS terminals 110 and 150 can be capped at 100%. If a specified number or percentage of POS terminals 110 and 150 causes the expected waiting time to exceed a predetermined threshold, the system can limit the number or percentage of POS terminals 110 and 150 to a range where the expected waiting time is below the predetermined threshold.

[0119] Service rate models can predict service rate capacity using hourly and store-level metrics. That is, service rate can be predicted based on hourly and store-level metric data and can be used to calculate key performance indicators for retail service locations 100, such as wait times, the number of waiting customers, and the utilization rates of POS terminals 110 and 150, in conjunction with modeled customer arrival rates 115. This model can be implemented using any suitable statistical procedure, such as extreme gradient boosting regression.

[0120] Service rate models can be derived from Figure 2 The machine learning module 232 described is implemented. As discussed above, a machine learning model can be trained using a dataset, such as one that can be stored in database 220, of past performance data from one or more retail service locations 100. For example, a portion of the past performance data can be used as a training dataset and another portion as a validation dataset to train the model. For instance, 70% of the past performance data can be used as a training dataset and the remaining 33% as a validation dataset. The service rate model can be trained using, for example, the number of items per transaction, the number or percentage of POS terminals in operation, seasonal periods (standard hours, Christmas and Thanksgiving), weekdays (Monday, Tuesday, etc.), peak periods (peak and off-peak), and the average number of transactions for each POS type at each retail service location 100 (attended POS terminal 110, unattended POS terminal 150, etc.). The service rate model can be, for example, a supervised machine learning model and can be validated based on mean absolute percentage error (MAPE). Furthermore, the root mean square error (RMSE) metric can be used instead of or in combination with MAPE to evaluate the training and validation of the service rate model. The service rate capacity of a retail service location is primarily affected by the number of items per transaction, the number or percentage of POS terminals in operation, and the average number of transactions.

[0121] The total service rate can be derived from the operating levels and service rates of POS terminals 110 and 150. Generally, the total service rate should be greater than the arrival rate of customer 115 to avoid indefinite waiting times for customer 115. During modeling, the arrival rate of customer 115 can be capped at a value less than the total service rate for each time period (e.g., per hour).

[0122] When retail location 100 removes one of POS terminals 110 or 150, customer 115 may seek different options to reduce their wait time. This movement or “traffic redistribution” between POS terminals 110 and 150 can depend on items 125 in customer 115’s shopping basket or cart 120 and the wait time of each of POS terminals 110 or 150. Based on the configuration of POS terminals 110 or 150 and a given wait time, an algorithm for predicting customer traffic redistribution can identify the availability of each of POS terminals 110 or 150. The algorithm can further redistribute traffic based on the transaction allocation pattern of retail location 100 and the availability of traffic allocation at POS terminals 110 or 150. Generally, reducing the number of POS terminals 110 or 150 results in longer wait times. Figure 5 A flowchart is depicted to model customer flow redistribution in a service location optimization approach based on one or more implementation schemes.

[0123] like Figure 5 As shown, the prediction of customer traffic redistribution can begin by receiving the configuration of POS terminals 110 or 150 in operation 510, such as a change in the configuration of a new retail service location 100 or a reconfiguration of an existing retail service location 100. In operation 520, the configuration of POS terminals 110 or 150 can be input into the customer traffic redistribution model. In operation 530, arrival rates and service rates can be predicted based on the previously trained model and the configuration of POS terminals 110 or 150. In operation 540, the average expected customer wait time can be calculated using the trained model and configuration of POS terminals 110 or 150. In operation 550, the calculated average expected customer wait time can be compared with a predetermined threshold (e.g., minutes or fractions of minutes). Different predetermined thresholds can be used for each type of POS terminal in POS terminals 110 and 150. If the calculated average expected customer wait time is less than a predetermined threshold, then in operation 560, it can be determined that there is no customer traffic redistribution. If the calculated average expected customer wait time is greater than or equal to the predetermined threshold, then in operation 570, the number of uncounted transactions is calculated, and the distribution of uncounted transactions in POS terminals 110 or 150 is determined based on, for example, the percentage of the total wait time in each of POS terminals 110 or 150. In operation 580, the modeled customer traffic redistribution can be updated to reflect the determined redistribution.

[0124] Retail service venue optimization process

[0125] By using the aforementioned metrics and models, the configuration and operation of retail service location 100 can be optimized in various aspects of its operation, including, for example, customer wait times (as described below). Figure 6 The average number of customers waiting in the queue (as discussed below) Figure 7 The average utilization rates of POS terminals 110 and 150 (as discussed below) Figure 8 (as discussed below), and labor costs (as discussed below). Figure 9 and Figure 10 (As discussed).

[0126] Figure 6 A flowchart is depicted in a method for optimizing service locations according to one or more implementation schemes, illustrating the calculation of optimized customer wait times. For example... Figure 6 As shown, in operation 605, the method can select data from one or more retail locations 100 from database 220 based on filters selected by the user. In operation 610, the method can receive optimization parameters from the user, such as a target average wait time per customer and the percentage of customer visits exceeding the target average wait time. The optimization parameters specified by the user can be, for example, hourly averages. The specified parameters can be normalized, for example, by taking the average of the recorded data for each retail location 100.

[0127] In operation 615, the method can calculate the total number of POS terminals 110 and 150 based on the percentage of POS terminals 110 and 150 in operation. This calculation can be based on using the average of the number of POS terminals 110 and 150 in operation as the percentage of POS terminals 110 and 150 in operation per hour. Furthermore, the method can predict customer arrival rates, for example, according to the methods discussed above. In operation 620, the method can iterate through the number of POS terminals 110 and 150 in operation, such as from 1 to 21, to predict the service rate of a given number of POS terminals 110 and 150 in operation. The number of POS terminals 110 and 150 in operation can be selected, for example, from data from retail service locations 100 and / or from other retail service locations 100. In operation 625, the method can calculate the predicted service rate of either POS terminals 110 or 150 in operation. In operation 630, the method can calculate the average waiting time in the queue of POS terminals 110 or 150 in operation. Such calculations may require that the arrival rate of customer 115 be less than the product of the predicted service rate and a specified number of operating POS terminals 110 and 150. To satisfy this condition, the arrival rate can be capped at less than the service rate, as discussed above. The average waiting time in the queue can be calculated based on the average number of customers waiting in the queue, calculated according to the percentage of customer visits exceeding a target average waiting time that can be specified by the user.

[0128] In operation 635, the method can divide the list of operating POS terminals 110 and 150 into a first list or data frame D1 of operating POS terminals 110 and 150 whose average waiting time is less than the target average waiting time per customer, and a second list or data frame D2 of operating POS terminals 110 and 150 whose average waiting time is greater than the target average waiting time per customer. In operation 635, the method can match the target average waiting time per customer with the first or second list of operating POS terminals 110 and 150 containing the operating POS terminals 110 and 150 whose average waiting time is closest to the target average waiting time per customer. If the first list of operating POS terminals 110 and 150 contains operating POS terminals 110 and 150 whose average waiting time is closest to the target average waiting time per customer, then in operation 650, the method can select the total number of operating POS terminals 110 and 150 from the first list as the optimal total number of POS terminals 110 and 150 under specified conditions. Otherwise, in operation 645, the method can select the total number of running POS terminals 110 and 150 from the second list as the optimal total number of POS terminals 110 and 150 under the specified conditions.

[0129] Figure 7 A flowchart is depicted illustrating the calculation of the optimized average number of customers waiting in a queue in a service location optimization method according to one or more implementation schemes. For example... Figure 7As shown, in operation 705, the method can select data from one or more retail locations 100 from database 220 based on a filter selected by the user. In operation 710, the method can receive optimization parameters from the user, such as the target average number of customers waiting in queues and the percentage of time periods exceeding the target average number of customers. The user's optimization parameters can be, for example, hourly averages. The specified parameters can be normalized, for example, by taking the average of the recorded data for each retail location 100. In operation 715, the method can calculate the total number of POS terminals 110 and 150 based on the percentage of POS terminals 110 and 150 in operation. This calculation can be based on using the average of the number of POS terminals 110 and 150 in operation as the hourly percentage of POS terminals 110 and 150 in operation. Furthermore, the method can predict customer arrival rates, for example, according to the methods discussed above. In operation 720, the method can iterate through the number of POS terminals 110 and 150 in operation, such as from 1 to 21, to predict the service rate for a given number of POS terminals 110 and 150 in operation. Operating POS terminals 110 and 150 can be selected, for example, from data from retail service location 100 and / or from other retail service locations 100. In operation 725, the method can calculate the predicted service rate of operating POS terminals 110 or 150. In operation 730, the method can calculate the average number of customers waiting in the queues of operating POS terminals 110 or 150. Such a calculation may require that the arrival rate of customers 115 is less than the product of the predicted service rate and a specified number of operating POS terminals 110 and 150. To satisfy this condition, the arrival rate can be capped at less than the service rate, as discussed above. The average number of customers waiting in the queues can be calculated based on a percentage of a time period exceeding, for example, a target average number of customers that can be specified by the user. In operation 735, the method can divide the list of operating POS terminals 110 and 150 into a first list or data frame D1 of operating POS terminals 110 and 150 where the average number of customers is less than the target average number of customers, and a second list or data frame D2 of operating POS terminals 110 and 150 where the average number of customers is greater than the target average number of customers. In operation 735, the method can match the target average number of customers with a first or second list of operating POS terminals 110 and 150 that contain the operating POS terminals 110 and 150 whose average number of customers is closest to the target average number of customers. If the first list of operating POS terminals 110 and 150 contains operating POS terminals 110 or 150 whose average number of customers is closest to the target average number of customers, then in operation 750, the method can select the total number of operating POS terminals 110 and 150 from the first list as the optimal total number of POS terminals 110 and 150 under the specified conditions.Otherwise, in operation 745, the method can select the total number of running POS terminals 110 and 150 from the second list as the optimal total number of POS terminals 110 and 150 under the specified conditions.

[0130] Figure 8 A flowchart is depicted illustrating the calculation of the optimized average utilization rate of POS terminals 110 and 150 in a service location optimization method according to one or more implementation schemes. For example... Figure 8As shown, in operation 805, the method can select data from one or more retail locations 100 from database 220 based on a filter selected by the user. In operation 810, the method can receive optimization parameters from the user, such as a target average utilization rate for POS terminals 110 and 150. The optimization parameters specified by the user can be, for example, hourly averages. The specified parameters can be normalized, for example, by taking the average of the recorded data for each retail location 100. In operation 815, the method can calculate the total number of POS terminals 110 and 150 based on the percentage of POS terminals 110 and 150 in operation. This calculation can be based on using the average of the number of POS terminals 110 and 150 in operation as the hourly percentage of POS terminals 110 and 150 in operation. Furthermore, the method can predict customer arrival rates, for example, according to the methods discussed above. In operation 820, the method can iterate through the number of POS terminals 110 and 150 in operation, such as from 1 to 21, to predict the service rate for a given number of POS terminals 110 and 150 in operation. Operating POS terminals 110 and 150 can be selected, for example, from data from retail service location 100 and / or from other retail service locations 100. In operation 825, the method can calculate the predicted service rate of operating POS terminals 110 or 150. In operation 830, the method can calculate the average utilization rate of POS terminals 110 or 150. Such a calculation may require that the arrival rate of customers 115 is less than the product of the predicted service rate and a specified number of operating POS terminals 110 and 150. To satisfy this condition, the arrival rate can be capped at less than the service rate, as discussed above. In operation 835, the method can divide the list of operating POS terminals 110 and 150 into a first list or data frame D1 of operating POS terminals 110 and 150 with an average utilization rate less than a target average utilization rate, and a second list or data frame D2 of operating POS terminals 110 and 150 with an average utilization rate greater than the target average utilization rate. In operation 835, the method can match the target average utilization rate with a first or second list of operating POS terminals 110 and 150 that contain operating POS terminals 110 and 150 whose average utilization rate is closest to the target average utilization rate. If the first list of operating POS terminals 110 and 150 contains operating POS terminals 110 and 150 whose average utilization rate is closest to the target average utilization rate, then in operation 850, the method can select the total number of operating POS terminals 110 and 150 from the first list as the optimal total number of POS terminals 110 and 150 under the specified conditions. Otherwise, in operation 845, the method can select the total number of operating POS terminals 110 and 150 from the second list as the optimal total number of POS terminals 110 and 150 under the specified conditions.

[0131] Figure 9A flowchart is depicted illustrating the calculation of optimized labor costs in a service location optimization method based on one or more implementation schemes. For example... Figure 9 As shown, in operation 905, the method can select data from one or more retail service locations 100 from database 220 based on filters selected by the user. In operation 910, the method can receive optimization parameters from the user, such as the total number of POS terminals 110 and 150, the percentage of operation of POS terminals 110 and 150, and hourly employee wages. For example, the user can specify a total of 10 POS terminals 110 and 150, a 50% operating percentage, and an hourly employee wage of $12.50. In operation 915, the method can calculate the total number of POS terminals 110 and 150 based on the percentage of operation of POS terminals 110 and 150. This calculation can be based on using the average of the operating POS terminals 110 and 150 as the hourly operating percentage of POS terminals 110 and 150. Furthermore, the method can predict customer arrival rates, for example, according to the methods discussed above. The method can then determine the starting and ending numbers of POS terminals 110 and 150 for calculation. For example, if the total number of operating POS terminals 110 and 150 is 10, and the percentage of operating POS terminals 110 and 150 is 50%, then the starting value could be 5 (10 - (10 * 50%)) and the ending value could be 15 (10 + (10 * 50%)). In operation 920, the method can cycle from the starting number to the ending number of operating POS terminals 110 and 150, such as from 5 to 15, to predict the service rate of a given number of operating POS terminals 110 and 150. Operating POS terminals 110 and 150 can be selected, for example, from data from retail service location 100 and / or from other retail service locations 100. In operation 925, the method can calculate the predicted service rate. In operation 930, the method can calculate the average waiting time in the queues of POS terminals 110 and 150. Such a calculation may require that the arrival rate of customer 115 is less than the product of the predicted service rate and the specified number of operating POS terminals 110 and 150. To meet this condition, the arrival rate can be capped at less than the service rate, as discussed above. In operation 935, the method can create a list with the total number of POS terminals 110 and 150, customer wait times, and labor costs, and can sort the list in descending order of the total number of POS terminals 110 and 150. In operation 935, the method can present this list in a user interface, for example, by drawing a line graph.

[0132] Figure 10A graph 1000 depicts the calculated optimized labor costs in a service location optimization method based on one or more implementation schemes. For this example, the user could specify a total of 10 POS terminals 110 and 150, a 50% operating percentage, and an employee wage of $12.50 per hour. Figure 10 As shown, for each number of POS terminals 110 and 150, data points 1040 can be plotted based on total service cost 1010 and average customer wait time 1030. Chart 1000 may also include reference lines 1030 indicating a specified number of users operating POS terminals 110 and 150.

[0133] Figure 11 A process flow 1100 describes a method for optimizing service locations according to one or more implementation schemes. For example... Figure 11 As shown, process flow 1100 may include calculating one or more models 1110, executing one or more algorithms 1120, and generating one or more outputs 1140. For example, the calculated model 1110 may include, for example, a clustering model for placing the retail service location 100 in one of several clusters similar to the retail service location 100 (such as the one mentioned above). Figure 4 The models discussed above include an arrival rate model for predicting the arrival rate of customers 115 at retail service location 100 (as discussed above) and a service rate model for predicting the service rate capacity of retail service location 100 (as discussed above).

[0134] Furthermore, one or more algorithms 1120 executed by process flow 1100 may include, for example, algorithms for predicting customer traffic redistribution (such as those mentioned above). Figure 5 The algorithms discussed include an algorithm for predicting the operational levels of POS terminals 110 and 150 (as discussed above), and one or more optimization algorithms performed by the retail service location optimizer 1130. The optimization algorithms performed by the retail service location optimizer 1130 may include, for example, an algorithm 1132 for calculating optimized customer wait times (as discussed above). Figure 6 The algorithm discussed above (1134) is used to calculate the optimized average number of customers waiting in the queue. Figure 7 The algorithm 1136 (discussed above) is used to calculate the optimized average utilization rate of POS terminals 110 and 150. Figure 8 (as discussed above), and algorithm 1138 for calculating optimized labor costs (as mentioned above). Figure 9 and Figure 10 (As discussed).

[0135] The generated outputs 1140 may include simulations 1142 of the operation and customer traffic of the retail service location 100, calculated optimizations of the operation of the retail service location 100, and various reports 1146 such as store metrics and basic data. For example, these outputs may be generated by the optimization module 235 and sent to the central office 270 or the back office 280, or may be generated by means of, as described above, the simulations 1140 and 2140. Figure 2 The report portal 290 discussed is available to any networked device, such as a mobile device application.

[0136] Figure 12 A flowchart is provided illustrating a method for optimizing service locations based on one or more implementation schemes. For example... Figure 12 As shown, in operation 1210, the method collects hourly data on store foot traffic and transactions. In operation 1220, the method collects store demographic data. In operation 1230, the method generates models for store clustering, customer arrival rate, and service rate capacity. In operation 1240, the method generates estimates of traffic redistribution and operational service location distribution. In operation 1250, the method receives user settings for service location optimization. In operation 1260, the method generates an optimized service location configuration. In operation 1270, the method generates a report on the optimized service location configuration.

[0137] Optimize user interface

[0138] As discussed above, service location optimization, according to one or more implementation schemes, may include, for example, through the methods described above regarding... Figure 2 The user interface module 250 discussed presents one or more user interfaces to control or report optimizations. Figure 14 and Figure 15 An exemplary user interface for service location optimization according to one or more implementation schemes is depicted.

[0139] like Figure 14 and Figure 15 As shown, the user interfaces 1400 and 1500 for service location optimization may include a parameter pane 1410 and a results pane 1440. The parameter pane 1410 may include a function selection control 1412 for the user to choose between a customization operation and an optimization operation. Figure 14 and Figure 15 As shown, the user has selected optimization function 1414. Based on a portion of optimization function 1414, parameter pane 1410 can display an additional user interface to control the optimization.

[0140] Once the optimization feature 1414 is selected, the parameter pane 1410 may also include an operational characteristic selection control 1416 for the user to select operational characteristics of the retail service location to be optimized. Operational characteristics of the retail service location may include, for example, the average wait time of customers in the queues of POS terminals, the average number of customers waiting in the queues of POS terminals, the average utilization rate of POS terminals, and the average service cost of POS terminals. Depending on the selected operational characteristics, the parameter pane 1410 may further display one or more parameters 1420 for controlling the optimization. Figure 14 As shown, users can select average service cost 1418 as the operating characteristic, and the displayed parameter 1420 may include the total number of POS terminals 1422, the percentage of POS terminals 1424 in operation, and average employee wages 1426. Figure 15 As shown, the user can select the average waiting time of customers in the queue of POS terminal 1518 as the running characteristic, and the displayed parameter 1420 may include a reference average performance metric of the target POS terminal, such as the average waiting time 1524, and the target percentage time 1522 that the reference average performance metric of the target POS terminal is satisfied.

[0141] Once the requested optimization is completed, the results pane 1440 can display the results of the requested optimization. For example, the results pane 1440 may include a text description 1442 of the requested optimization, such as "Service Costs and Waiting Times". Furthermore, the results pane 1440 may include detailed results of the requested optimization. For example, such as... Figure 14 As shown, a graph such as wait time versus service cost can be displayed (1444). Or, as... Figure 15 As shown, this may display the number of POS terminals required to meet the specified reference average performance metric, which is 1544.

[0142] Whenever a user updates the settings in the runtime feature selection control 1416 or parameter 1420, the results pane 1440 can be updated automatically, allowing for quick and intuitive exploration of the optimization parameters for the retail service location 100.

[0143] computing devices

[0144] Figure 13 The exemplary embodiments according to this disclosure can be configured to perform Figures 1 to 12A simplified functional block diagram of a computer 1300 as an environment and / or method of the device. For example, device 1300 may include a central processing unit (CPU) 1320. CPU 1320 may be any type of processor device, including, for example, any type of dedicated or general-purpose microprocessor device. As those skilled in the art will understand, CPU 1320 may also be a single processor in a multi-core / multi-processor system (such a system operating independently), or a single processor in a cluster of computing devices operating in a cluster or server group. CPU 1320 may be connected to data communication infrastructure 1310, such as a bus, message queue, network, or multi-core messaging scheme.

[0145] Device 1300 may also include main memory 1340, such as random access memory (RAM), and may also include secondary memory 1330. Secondary memory 1330 (e.g., read-only memory (ROM)) may be, for example, a hard disk drive or a removable storage drive. Such a removable storage drive may include, for example, a floppy disk drive, a magnetic tape drive, an optical disc drive, flash memory, etc. The removable storage drive in this example reads from and / or writes to the removable storage unit in a well-known manner. The removable storage unit may include floppy disks, magnetic tapes, optical discs, etc., read and written by the removable storage drive. As those skilled in the art will understand, such a removable storage unit typically includes a computer-usable storage medium storing computer software and / or data.

[0146] In an alternative implementation, secondary storage 1330 may include other similar means for allowing computer programs or other instructions to be loaded into device 1300. Examples of such means may include a program box and box interface (such as a program box and box interface in a video game device), a removable storage chip (such as an EPROM or PROM) and associated test sockets, as well as other removable storage units and interfaces that allow software and data to be transferred from the removable storage unit to device 1300.

[0147] Device 1300 may also include a communication interface (“COM”) 1360. Communication interface 1360 allows the transfer of software and data between device 1300 and external devices. Communication interface 1360 may include a modem, network interface (such as an Ethernet card), communication port, PCMCIA slot, and card, etc. Software and data transferred via communication interface 1360 may be in the form of signals, which may be electronic, electromagnetic, optical, or other signals that can be received by communication interface 1360. These signals may be provided to communication interface 1360 via communication paths of device 1300, which may be implemented using, for example, wires or cables, optical fibers, telephone lines, cellular telephone links, RF links, or other communication channels.

[0148] The hardware components, operating system, and programming language of such a device are essentially conventional and are assumed to be sufficiently familiar to those skilled in the art. Device 1300 may also include input and output ports 1350 for connection to input and output devices such as keyboards, mice, touchscreens, monitors, displays, etc. Of course, various server functions can be implemented in a distributed manner on multiple similar platforms to distribute the workload. Alternatively, the server can be implemented through appropriate programming of a single computer hardware platform.

[0149] The program aspect of this technology can be considered a "product" or "artifact," typically in the form of executable code and / or associated data, carried or contained in a type of machine-readable medium. "Storage" type media includes any or all tangible memory, or associated modules, of computers, processors, etc., such as various semiconductor memories, tape drives, disk drives, etc., which can provide non-transitory storage for software programming at any time. Sometimes, all or part of the software may be transmitted via the Internet or various other telecommunications networks. For example, such communication enables the loading of software from one computer or processor to another, such as from a management server or host computer of a mobile communication network to a server's computer platform and / or from a server to a mobile device. Therefore, another type of medium that can carry software elements includes light waves, radio waves, and electromagnetic waves used on physical interfaces between local devices via wired and optical ground networks and various air links. Physical elements carrying such waves, such as wired or wireless links, optical links, etc., can also be considered as media carrying software. As used herein, unless limited to non-transitory tangible "storage" media, the term "readable medium" for a computer or machine refers to any medium that participates in providing instructions to a processor for execution.

[0150] References to any particular activity are provided in this disclosure for convenience only and not for limitation. Those skilled in the art will recognize that the concepts upon which the disclosed apparatus and methods are based can be used in any suitable activity. This disclosure can be understood with reference to the following description and accompanying drawings, wherein like elements are indicated by like reference numerals.

[0151] The terminology used above may be interpreted in its broadest and most reasonable manner, even when used in conjunction with the detailed description of certain specific examples of this disclosure. In fact, some terms may even be emphasized above; however, any term intended to be interpreted in any limiting manner will be disclosed and specifically defined in this Detailed Description section. The general description and detailed description are exemplary and explanatory only and do not limit the claimed features.

[0152] In this disclosure, the term "based on" means "at least partially based on". Unless the context otherwise requires, the singular forms "a", "an", and "the" include the plural objects. The term "exemplary" is used in the sense of "example" rather than "ideal". The terms "comprises", "comprising", "includes", "including", or other variations thereof are intended to cover non-exclusive inclusion, so a process, method, or product that includes a list of elements does not necessarily include only those elements, but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. The term "or" is used separately such that "at least one of A or B" includes (A), (B), (A and A), (A and B), etc. Related terms such as "substantially" and "usually" are used to indicate possible variations of ±10% in the specified or understood value.

[0153] As used herein, terms such as “user” typically encompass pet parents and / or multiple pet parents. Terms such as “pet” typically encompass a user’s pets, where the term may cover multiple pets. Furthermore, the term “pet” refers to any type of animal, including domestic animals. Terms such as “provider” typically encompass pet care businesses.

[0154] It should be understood that in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped in a single embodiment, drawing, or description thereof in order to simplify the disclosure and aid in understanding one or more of the various inventive aspects. However, this method of disclosure should not be construed as reflecting an intention that the claimed invention requires more features than expressly recited in each claim. More precisely, as reflected in the appended claims, the inventive aspect lies in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into this detailed description, wherein each claim is independently a separate embodiment of the invention.

[0155] Furthermore, while some embodiments described herein include some but not others of the features included in other embodiments, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments, as will be understood by those skilled in the art. For example, any embodiment of the claimed embodiments in the appended claims may be used in any combination.

[0156] Therefore, although certain embodiments have been described, those skilled in the art will recognize that other and further modifications can be made to these embodiments without departing from the spirit of the invention, and it is intended to claim protection for all such changes and modifications falling within the scope of the invention. For example, functionality can be added or removed from the block diagrams, and operations can be interchanged within functional blocks. Steps can be added or removed from the described methods within the scope of the invention.

[0157] The subject matter disclosed above should be considered illustrative rather than restrictive, and the appended claims are intended to cover all such modifications, enhancements, and other embodiments falling within the true spirit and scope of this disclosure. Therefore, to the maximum extent permitted by law, the scope of this disclosure will be determined by the broadest permissible interpretation of the appended claims and their equivalents, and should not be bound or limited by the foregoing detailed description. While various embodiments of this disclosure have been described, it will be apparent to those skilled in the art that further embodiments are possible within the scope of this disclosure. Therefore, this disclosure is not limited except as provided in the appended claims and their equivalents.

Claims

1. A computer-implemented method for optimizing the operational characteristics of a service location, the method comprising: Receive past performance data for one or more operational characteristics of one or more reference service locations; The past performance data includes customer transactions from multiple operating POS terminals; Receive target values ​​from the user for additional operational characteristics of the target service location; Receive operational data from one or more operating point-of-sale (POS) terminals at the target service location and / or receive operational data from the reference service location; Use the received past performance data to train the machine learning model; Use trained machine learning models to calculate predicted service rates or predicted arrival rates; Based on the predicted service rate or predicted arrival rate, one or more customer traffic flows are redistributed among the plurality of operating POS terminals; After the reallocation, the average value of the additional operating characteristics of each of the plurality of operating POS terminals is calculated; The average POS terminal is determined as the POS terminal among the one or more operating POS terminals by using the calculated average value of the additional operating characteristics that matches the target value used for the additional operating characteristics; The average POS terminal was selected as the representative POS terminal. The operating characteristics of the target POS terminal at the target service location are set to be equal to the operating characteristics of the representative POS terminal; and The machine learning model is updated using the newly received operational characteristics for the plurality of operating POS terminals.

2. The computer-implemented method of claim 1, wherein the additional operational characteristic is one of the following: the average waiting time of customers in the queue of the target POS terminal, the average number of customers waiting in the queue of the target POS terminal, the average utilization rate of the target POS terminal, and the average service cost of the target POS terminal.

3. The computer-implemented method as described in claim 1, wherein selecting the POS terminal among the one or more operating POS terminals further includes: The one or more operating POS terminals are divided into a first list of operating POS terminals having values ​​of the additional operating characteristics that are less than the target value for the additional operating characteristics, and a second list of operating POS terminals having values ​​of the additional operating characteristics that are greater than or equal to the target value for the additional operating characteristics. as well as One of the first list and the second list is selected as the representative list of POS terminals, the representative list of POS terminals containing operating POS terminals whose calculated average value of the additional operating characteristics matches the target value used for the additional operating characteristics.

Images