Scheduling method, device and equipment for bank cloud counter teller and computer program product

Abstract

The invention discloses a bank cloud counter teller scheduling method and device, electronic equipment and a computer program product. The method comprises the steps of obtaining historical customer arrival data; according to the historical customer arrival data, a customer arrival quantity prediction algorithm is utilized to determine a customer arrival quantity estimated value in a future time period; based on a queuing theory model, an integer programming problem is constructed according to the customer arrival quantity estimated value and cloud counter teller service data; the integer programming problem is converted into a QUBO model; and solving the QUBO model by using a quantum solving algorithm to obtain the number of tellers in each scheduling time period in the future. According to the method, a cloud counter seat scheduling problem is abstracted into a nonlinear programming problem based on a queuing theory model, a QUBO model and a quantum computing technology are applied to bank cloud teller scheduling optimization for solving the programming problem, the powerful computing power of the quantum computing technology is utilized, the problem that a classical algorithm falls into a local optimal solution is avoided, and the scheduling efficiency is improved. And a high-quality solution is provided for an integer programming problem of NP-difficulty.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and in particular to a method, device, electronic equipment, and computer program product for scheduling tellers at a bank cloud counter. Background Art

[0002] At present, although the remote counter (i.e. cloud counter) system of commercial banks is in a golden period of vigorous development, there is still a lack of an efficient and innovative remote counter scheduling solution. Most of the existing technical solutions focus on the scheduling management of counter seats at outlets. These solutions mainly formulate scheduling plans based on the customer flow of specific areas or outlets and the actual situation of teller resources. However, the challenges faced by remote counter scheduling are more complex. It needs to comprehensively consider the customer call demand of the commercial bank's entire online channel and the service carrying capacity of the remote bank. The breadth and depth of its data integration far exceed the scheduling of counter seats at outlets.

[0003] Among the existing scheduling methods, many solutions still rely too much on the personal experience of experts and traditional classical algorithms. This reliance not only limits the flexibility of scheduling strategies, but also makes the system incapable of dealing with complex and changing business scenarios. In addition, when it comes to solving scheduling planning problems, existing classical solution algorithms are often prone to falling into the dilemma of local optimal solutions, especially when dealing with NP-hard integer programming problems. This problem is particularly prominent. Summary of the invention

[0004] The embodiments of the present application provide a method, device, electronic equipment, and computer program product for scheduling tellers at a bank cloud counter to achieve efficient scheduling and intelligent operation of remote counter services.

[0005] The present application embodiment adopts the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for scheduling a teller at a bank cloud counter, the method comprising:

[0007] Get historical customer arrival data;

[0008] Determine an estimated value of the number of customers arriving in a future time period using a customer arrival number prediction algorithm based on the historical customer arrival data;

[0009] Based on the queuing theory model, an integer programming problem for cloud counter teller scheduling is constructed according to the estimated number of customers arriving in the future time period and the service data of cloud counter tellers;

[0010] Convert the integer programming problem of the cloud counter teller scheduling into a QUBO model;

[0011] The QUBO model is solved using a preset quantum solution algorithm to obtain the number of cloud counter tellers in each future scheduling time period.

[0012] Optionally, converting the integer programming problem of the cloud counter teller scheduling into a QUBO model includes:

[0013] Define binary decision variables;

[0014] The QUBO model is constructed according to the binary decision variables and the integer programming problem of the cloud counter teller scheduling.

[0015] Optionally, constructing the QUBO model according to the binary decision variables and the integer programming problem of the cloud counter teller scheduling includes:

[0016] Constructing the loss function of the QUBO model according to the binary decision variables and the loss function of the integer programming problem of the cloud counter teller scheduling;

[0017] Constructing the constraint condition of the binary decision variable, wherein the constraint condition of the binary decision variable is configured such that one scheduling time period corresponds to one number of tellers;

[0018] The loss function of the QUBO model is adjusted according to the constraint conditions of the binary decision variables to obtain an adjusted loss function of the QUBO model.

[0019] Optionally, the QUBO model is solved by using a preset quantum solution algorithm to obtain the number of cloud counter tellers in each future scheduling time period, including:

[0020] The QUBO model is solved using a quantum solution algorithm of a coherent Ising machine or a quantum annealing machine to obtain the number of cloud counter tellers in each future scheduling time period.

[0021] Optionally, the customer arrival number prediction algorithm includes a time series autoregressive model, and determining the estimated value of the customer arrival number in a future time period by using the customer arrival number prediction algorithm according to the historical customer arrival data includes:

[0022] Inputting the historical customer arrival data into the time series autoregressive model to obtain a probability distribution of the number of customer arrivals per day in the future output by the time series autoregressive model;

[0023] Average the probability distribution of the number of customers arriving each day in the future to get the number of customers arriving each day in the future;

[0024] Perform statistical analysis on the historical customer arrival data to calculate the ratio of the number of customer arrivals in each time period of each day in the future to the total number of customer arrivals for the whole day;

[0025] Based on the number of customer arrivals each day in the future and the proportion of the number of customer arrivals in each time period of each day to the total number of customer arrivals for the day, calculate the estimated number of customer arrivals in each time period of each day in the future.

[0026] Optionally, the integer programming problem of cloud counter teller scheduling based on the queuing theory model and according to the estimated value of the number of customers arriving in the future time period and the service data of the cloud counter tellers includes:

[0027] According to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers, the functional relationship between the number of tellers and the average waiting time of customers is constructed based on the queuing theory model;

[0028] An integer programming problem for cloud counter teller scheduling is constructed based on the functional relationship between the number of tellers and the average waiting time of customers.

[0029] Optionally, constructing a functional relationship between the number of tellers and the average waiting time of customers based on a queuing theory model according to the estimated number of customers arriving in the future time period and the service data of the tellers at the cloud counter includes:

[0030] Performing statistical analysis on the historical customer arrival data to determine the distribution characteristics of the number of customer incoming calls, the time interval between customer incoming calls, and the teller service time;

[0031] Determine the target queuing theory model based on the distribution characteristics of the number of customer calls, the time interval between customer calls, and the teller service time;

[0032] Determine the average number of incoming customer calls per second in each period based on the estimated number of customer arrivals in each period of the future day;

[0033] Performing statistical analysis on the historical customer arrival data to calculate the mean of teller service time;

[0034] According to the average number of incoming customer calls per second and the mean of teller service time in each time period, the functional relationship between the number of tellers and the average waiting time of customers is constructed based on the target queuing theory model.

[0035] In a second aspect, an embodiment of the present application further provides a scheduling device for a teller at a bank cloud counter, wherein the scheduling device for a teller at a bank cloud counter comprises:

[0036] An acquisition unit, used to acquire historical customer arrival data;

[0037] A determination unit, configured to determine an estimated value of the number of customers arriving in a future time period based on the historical customer arrival data and using a customer arrival number prediction algorithm;

[0038] A construction unit, for constructing an integer programming problem for scheduling cloud counter tellers based on a queuing theory model and according to an estimated number of customers arriving in a future time period and service data of cloud counter tellers;

[0039] A conversion unit, used for converting the integer programming problem of the cloud counter teller scheduling into a QUBO model;

[0040] The solving unit is used to solve the QUBO model by using a preset quantum solving algorithm to obtain the number of cloud counter tellers in each future scheduling time period.

[0041] In a third aspect, an embodiment of the present application further provides a device, including:

[0042] A processor; and a memory arranged to store computer executable instructions, which, when executed, cause the processor to execute any of the aforementioned methods for scheduling tellers at a bank cloud counter.

[0043] In a fourth aspect, an embodiment of the present application further provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements any of the aforementioned methods for scheduling tellers at a bank cloud counter.

[0044] At least one of the above-mentioned technical solutions adopted in the embodiments of the present application can achieve the following beneficial effects: the scheduling method of bank cloud counter tellers in the embodiments of the present application first obtains historical customer arrival data; then, based on the historical customer arrival data, a customer arrival number prediction algorithm is used to determine the estimated number of customer arrivals in the future time period; then, based on the queuing theory model, an integer programming problem for cloud counter teller scheduling is constructed according to the estimated number of customer arrivals in the future time period and the service data of the cloud counter tellers; then, the integer programming problem for cloud counter teller scheduling is converted into a QUBO model; finally, the QUBO model is solved using a preset quantum solution algorithm to obtain the number of cloud counter tellers in each future scheduling time period. The scheduling method for bank cloud counter tellers in the embodiment of the present application organizes data and mines features based on the traffic characteristics of cloud counters and the response status of remote counter services of all commercial banks. The remote counter seat scheduling problem is abstracted as a nonlinear programming problem based on the queuing theory model, and feasible combination schemes are calculated. Compared with simple machine learning and neural network schemes, it is more suitable for queue problems, has strong interpretability, and has higher accuracy of calculation results. For solving the planning problem, the QUBO model and quantum computing technology are applied to the scheduling optimization of bank cloud tellers, and the powerful computing power of quantum computing technology is used to search the solution space in the entire domain, thereby avoiding the situation where the classical algorithm falls into the local optimal solution, and providing a high-quality solution for the NP-hard integer programming problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0046] Figure 1 A flow chart of a method for scheduling tellers at a bank cloud counter in an embodiment of the present application;

[0047] Figure 2 A frequency statistical distribution diagram of an incoming call time interval in an embodiment of the present application;

[0048] Figure 3 This is a structural schematic diagram of a scheduling device for a bank cloud counter teller in an embodiment of the present application;

[0049] Figure 4 This is a schematic diagram of the structure of a device in an embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present application clearer, the technical solution of the present application will be clearly and completely described below in combination with the specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application.

[0051] The technical solutions provided by various embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0052] The present application embodiment provides a method for scheduling tellers at a bank cloud counter, such as Figure 1 As shown, a flow chart of a method for scheduling a teller at a bank cloud counter in an embodiment of the present application is provided. The method for scheduling a teller at a bank cloud counter includes at least the following steps S110 to S150:

[0053] Step S110, obtaining historical customer arrival data.

[0054] When scheduling tellers at a bank cloud counter, the embodiment of the present application needs to first obtain customer arrival data for a period of time in the past (such as the past three years) from the bank's database. Customer arrival data refers to data on customers calling into the bank cloud counter. These data may include key information such as the customer's arrival time and arrival date, which serves as the basis for subsequent predictions and analysis.

[0055] Step S120, based on the historical customer arrival data, using a customer arrival quantity prediction algorithm to determine an estimated value of the customer arrival quantity in a future time period.

[0056] After obtaining the historical customer arrival data, it is necessary to further use the customer arrival number prediction algorithm to predict the estimated value of the customer arrival number in the future. The customer arrival number prediction algorithm of the embodiment of the present application can be a prediction model obtained by training neural networks such as existing recursive neural networks and recurrent neural networks, or it can be a traditional statistical analysis prediction algorithm. The specific prediction algorithm to be used can be flexibly selected by those skilled in the art according to actual needs, and no specific limitation is made here.

[0057] Step S130, based on the queuing theory model, an integer programming problem for scheduling cloud counter tellers is constructed according to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers.

[0058] Queuing theory is a mathematical theory and method for studying queuing phenomena in service systems. In the cloud counter teller scheduling scenario, parameters such as the number of tellers, customer waiting time, and their relationships meet the application premise of the queuing theory model. Integer programming is a special mathematical programming problem that requires the value of the variable to be an integer. In the cloud counter teller scheduling scenario, the goal of cloud counter teller scheduling is to find the optimal number of tellers that meets business requirements. Therefore, this problem belongs to an integer programming problem.

[0059] Based on this, the embodiment of the present application constructs an integer programming problem for cloud counter teller scheduling based on a queuing theory model, combined with the estimated number of customer arrivals in the future and the service data of cloud counter tellers (such as teller service time and other data).

[0060] Step S140, converting the integer programming problem of the cloud counter teller scheduling into a QUBO model.

[0061] To solve the integer programming problem constructed in the previous steps, it is necessary to first convert the integer programming problem into a QUBO (Quadratic Unconstrained Binary Optimization) model. The QUBO model is a specific form of mathematical modeling, mainly used to solve combinatorial optimization problems. It binarizes the variables in the original problem, that is, the variables can only take 0 or 1 (or -1 and 1, depending on whether the domain space used is Boolean space or Ising space), and constructs a quadratic objective function to minimize or maximize the function value.

[0062] Converting integer programming problems into QUBO models can simplify the solution process and reduce the difficulty of solution.

[0063] Step S150, using a preset quantum solution algorithm to solve the QUBO model to obtain the number of cloud counter tellers in each future scheduling time period.

[0064] After converting the integer programming problem into a QUBO model, the embodiment of the present application needs to use quantum solving algorithms such as coherent Ising machines or quantum annealing machines and other quantum computing technologies to find the optimal solution, so as to obtain the optimal number of tellers in each future scheduling time period.

[0065] The scheduling method for bank cloud counter tellers in the embodiment of the present application organizes data and mines features based on the traffic characteristics of cloud counters and the response status of remote counter services of all commercial banks. The remote counter seat scheduling problem is abstracted as a nonlinear programming problem based on the queuing theory model, and feasible combination schemes are calculated. Compared with simple machine learning and neural network schemes, it is more suitable for queue problems, has strong interpretability, and has higher accuracy of calculation results. For solving the planning problem, the QUBO model and quantum computing technology are applied to the scheduling optimization of bank cloud tellers, and the powerful computing power of quantum computing technology is used to search the solution space in the entire domain, thereby avoiding the situation where the classical algorithm falls into the local optimal solution, and providing a high-quality solution for the NP-hard integer programming problem.

[0066] In some embodiments of the present application, the customer arrival number prediction algorithm includes a time series autoregressive model, and the use of the customer arrival number prediction algorithm to determine the estimated value of the customer arrival number in the future time period based on the historical customer arrival data includes: inputting the historical customer arrival data into the time series autoregressive model to obtain the probability distribution of the future daily customer arrival number output by the time series autoregressive model; averaging the probability distribution of the future daily customer arrival number to obtain the future daily customer arrival number; performing statistical analysis on the historical customer arrival data to calculate the proportion of the number of customer arrivals in each time period of each day in the future to the total number of customer arrivals for the day; and calculating the estimated value of the number of customer arrivals in each time period of each day in the future based on the number of customer arrivals in the future and the proportion of the number of customer arrivals in each time period of each day to the total number of customer arrivals for the day.

[0067] The estimated value of the number of customers arriving in the future time period defined in the embodiment of the present application may refer to the estimated value of the number of customers arriving in each time period in the future. The estimation of the number of customers arriving in each time period in the future may be implemented in the following two aspects:

[0068] On the one hand, by processing the historical customer arrival data using the customer arrival number prediction algorithm, the daily customer arrival number in the future time period can be predicted as result A. That is, the customer arrival number predicted by the customer arrival number prediction algorithm is a prediction result of the daily dimension.

[0069] On the other hand, by conducting statistical analysis on the historical data of daily calls received by the cloud cabinet system, we can calculate the ratio of the number of customers calling into the cloud cabinet system in each period to the total number of customers for the whole day. Using samples over a period of time, such as the past year, we can calculate the average ratio of the number of customers connected in each period to the total number of customers for the whole day, and obtain result B.

[0070] Based on the above two aspects, the customer arrival number prediction algorithm can achieve a relatively coarse-grained prediction of the future customer arrival number. The statistical analysis algorithm is used to perform statistical analysis on historical data, and it can be found that the proportion of customer inbound calls in each period of each day is relatively stable. For example, the busy period usually occurs between 9:30 and 10:30 every day, and the number of inbound customers accounts for about 8.9% of the whole day. In this way, the coarse-grained future customer arrival number and the fine-grained customer inbound call proportion obtained by statistical analysis, that is, using result A*result B, can obtain the estimated value of customer arrival number in each period of each day in the future.

[0071] Compared with the method of directly training the model to predict the fine-grained number of customer arrivals, the above implementation method can greatly reduce the data requirements for model training. By combining a simple statistical analysis algorithm, it fully utilizes the relatively stable proportion of customer incoming calls and improves the accuracy and efficiency of the prediction.

[0072] Furthermore, for the prediction of future data in the first aspect mentioned above, the current mainstream machine learning methods are neural networks such as recursive neural networks (RNN) and long short-term memory networks (LSTM). Unlike ordinary feedforward neural networks, recursive neural networks can use time series to analyze inputs. That is, ordinary neural networks will think that the content input at time t is completely unrelated to the content input at time t+1, while recursive neural networks will treat the output value x at time t-1 as the value of the input at time t+1. t-1 is copied to time t, and the input x at time t t After integration, the output is formed after passing through an activation function with weights and biases. Similarly, LSTM is based on recurrent neural networks and adds gate control to solve long-term dependency problems.

[0073] The customer arrival quantity prediction algorithm of the embodiment of the present application can be a time series autoregression model. The time series autoregression model is similar to the above-mentioned scheme, but the difference is that the time series autoregression model is obtained based on recurrent neural network training. It does not simply output a certain prediction value directly, but outputs a probability distribution of the prediction value.

[0074] When using the time series autoregression model to predict the number of customer arrivals in the future, the historical customer arrival data can be preprocessed first. For example, the historical customer arrival data can be converted into time series data and subjected to sliding window processing. The data after sliding window processing can then be used as the input of the time series autoregression model, so that the time series autoregression model can predict the number of customer arrivals in the future.

[0075] Specifically, the time series autoregression model is used to perform autoregressive prediction on the probability parameters of the time series of the number of customers per day, and the probability distribution of the predicted number of customers arriving in the next X days is output and the confidence interval of the prediction is given on this basis. The average value of the probability distribution is taken as the number of customers arriving in the next X days. Repeat the previous step Y times, and roll-forward the time series of the number of customers arriving in the next X*Y days to obtain the predicted number of customers arriving every day in the future.

[0076] The probability distribution of the predicted value output by the time series autoregressive model adopted in the embodiment of the present application, on the one hand, takes into account that many processes themselves have random properties, and therefore the output probability distribution is closer to the essence and the prediction accuracy will be higher; on the other hand, compared with traditional neural networks, the uncertainty of the prediction and related risks can be evaluated, thereby further improving the prediction accuracy.

[0077] In some embodiments of the present application, the time series autoregressive model can be trained in the following manner: the historical customer arrival data is processed, the number of customers arriving at the cloud counter every day is counted, the number of customers arriving every day is converted into time series data, and sliding window processing is performed to convert all the time series data into sliding window data, wherein the size of the sliding window is set to 30 days, for example, and the step length is set to 5 days, for example. The sliding window data is divided into a training set, a validation set, and a test set in a ratio of 3:3:1 in order as model training test data.

[0078] After preparing the model training test data, call the time series autoregressive model interface, and the time series autoregressive model requests the required data from the historical database for training model parameters, obtains the neural network weights, and solidifies the model weights. During the model training process, the automatic hyperparameter search technology can be used to find the best hyperparameters of the model through the Bayesian optimization algorithm to minimize the model loss function loss of the verification test set, and finally obtain a time series autoregressive model that meets the use requirements.

[0079] In some embodiments of the present application, the integer programming problem of scheduling cloud counter tellers is constructed based on the queuing theory model according to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers, including: constructing a functional relationship between the number of tellers and the average waiting time of customers based on the queuing theory model according to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers; constructing an integer programming problem of scheduling cloud counter tellers according to the functional relationship between the number of tellers and the average waiting time of customers.

[0080] After obtaining the estimated number of customers arriving in the future, it is necessary to use this data and the service data of the cloud counter tellers (such as the service time of each teller, etc.) to build a functional relationship between the number of tellers and the average waiting time of customers. This functional relationship needs to be built based on the queuing theory model. In this step, by considering the number of incoming customer calls and the average waiting time of customers, the queuing theory model is used to calculate the average waiting time of customers under different numbers of tellers, thereby building a functional relationship between the number of tellers and the average waiting time of customers.

[0081] After obtaining the functional relationship between the number of tellers and the average waiting time of customers, it is necessary to construct an integer programming problem for cloud counter teller scheduling based on this functional relationship. In this step, a loss function (such as the minimum number of tellers corresponding to the average waiting time of customers who are satisfied with the business) and a series of constraints (such as the upper limit of the number of tellers, the limit of customer waiting time, etc.) can be constructed based on the functional relationship between the number of tellers and the average waiting time of customers, and then an integer programming problem can be constructed based on these conditions. The solution to this problem is the number of cloud counter tellers in each future scheduling time period.

[0082] In some embodiments of the present application, constructing the functional relationship between the number of tellers and the average waiting time of customers based on the queuing theory model according to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers includes: performing statistical analysis on the historical customer arrival data to determine the distribution characteristics of the number of customer incoming calls, the time interval of customer incoming calls and the teller service time; determining the target queuing theory model according to the distribution characteristics of the number of customer incoming calls, the time interval of customer incoming calls and the teller service time; determining the average number of customer incoming calls per second in each time period according to the estimated number of customers arriving in each time period every day in the future; performing statistical analysis on the historical customer arrival data to calculate the mean of the teller service time; constructing the functional relationship between the number of tellers and the average waiting time of customers based on the target queuing theory model according to the average number of customer incoming calls per second and the mean of the teller service time in each time period.

[0083] Traditional queuing theory provides a mathematical model for service-type queues. In the embodiment of the present application, the cloud counter service conforms to the M / G / C model, where M represents the number of incoming customer calls and conforms to the Poisson distribution. It can be seen that the interval between customer incoming calls obeys the exponential distribution. G represents the general distribution. In the embodiment of the present application, it can be determined from the data statistical dimension that the teller service time conforms to the k-order Erlang distribution. C represents the number of counters. In the embodiment of the present application, C represents the number of tellers.

[0084] The distribution characteristics of the two random variables, customer call time interval and teller service time, can be obtained through historical data statistics. Taking the call time interval as an example, a frequency statistical distribution graph with a bin interval of 1 second in each group as the x-axis is drawn. By observing the frequency statistical distribution graph, it is found that the distribution of the random variable time interval distribution conforms to the exponential distribution. Figure 2 As shown, a frequency statistical distribution diagram of the incoming call time interval in an embodiment of the present application is provided.

[0085] There are two methods to detect the distribution. One method is to compare the sample data distribution and the QQ graph of the theoretical distribution. If the sample point is close to the theoretical value, it means that it conforms to the distribution. Another method is to use a non-parametric test to compare the sample distribution with the theoretical distribution through statistical judgment. Of course, the specific verification method can be flexibly selected by those skilled in the art according to actual needs, and no specific limitation is made here.

[0086] The verification of the time interval of customer incoming calls is consistent with the above statistical detection method, indicating that the use of the queuing theory model M / G / C is in line with the actual situation. Therefore, we can combine the estimated number of customers arriving in the future and the service time of cloud counter tellers and other data to build a functional relationship between the number of tellers and the average waiting time of customers based on the M / G / C model.

[0087] In some embodiments of the present application, the service data of the cloud counter teller includes the teller service time, and the functional relationship between the number of tellers and the average waiting time of customers is constructed based on the target queuing theory model according to the estimated number of customers arriving in the future time period and the service data of the cloud counter teller, including: determining the average number of customer incoming calls per second in each time period according to the estimated number of customers arriving in each time period every day in the future; performing statistical analysis on the historical customer arrival data to calculate the mean of the teller service time; and constructing the functional relationship between the number of tellers and the average waiting time of customers based on the target queuing theory model according to the average number of customer incoming calls per second in each time period and the mean of the teller service time.

[0088] When constructing the functional relationship between the number of tellers and the average waiting time of customers based on the M / G / C model, we first divide the estimated number of customers arriving in each period by the time, that is, the average number of customer calls per second in each period, to ensure the uniformity of subsequent calculation standards. Assume that this value is λ, that is, the customer call time interval obeys the coefficient In addition, the mean value μ of the teller’s service time can also be calculated based on historical data.

[0089] According to the calculation formula of queuing theory, the average waiting time of customers and the number of tellers form a functional relationship, which is as follows:

[0090]

[0091] Among them, υ is the coefficient of variation of teller service time, L q is the average queue length, L s is the number of customers in the system, ρ i is the service intensity of the cloud counter at the i-th moment, n i is the number of counters (i.e., the number of tellers) at the i-th moment, λ i is the arrival rate at the i-th moment (i.e., the number of customers arriving per unit time), μ i is the service rate at the i-th moment (i.e., the mean service time of the teller), is the average waiting time of the M / M / r queuing system with various parameters substituted into the i-th moment.

[0092] By substituting the number of tellers into the above formula, we can get the number of tellers (n i ) and the average customer waiting time (t i )’s function value relationship.

[0093] In some embodiments of the present application, constructing the integer programming problem of cloud counter teller scheduling based on the functional relationship between the number of tellers and the average waiting time of customers includes: constructing a loss function of the integer programming problem of cloud counter teller scheduling based on the number of tellers and the average waiting time of customers; constructing multiple constraints of the integer programming problem of cloud counter teller scheduling based on the functional relationship between the number of tellers and the average waiting time of customers, the daily upper limit of the number of tellers, and the average waiting time of customers who are satisfied with the service.

[0094] Arranging the number of tellers for each time period of the day will evolve into an integer programming problem. In this problem, the variable can be, for example, the number of working hours allocated to each of the 21 time slots of a day (such as working hours 8:30-19:00, each half hour is a slot), that is, the number of tellers.

[0095] First, construct the loss function of the integer programming problem, which can be expressed as follows:

[0096]

[0097] Among them, a is the penalty coefficient for the number of tellers, and b is the penalty coefficient for customer waiting time, which can be customized by the business party.

[0098] Then, based on the upper limit of the number of tellers per day, the following constraints can be further constructed:

[0099] n i ≤N, (8)

[0100]

[0101] n i ·μ i ≥λ i , (10)

[0102]

[0103] t i ≤T, (12)

[0104] Among them, N is the upper limit of the number of tellers per day, n i is the number of tellers assigned to each time slot. Therefore, the constraint of formula (8) is that the number of tellers assigned to each time slot cannot exceed the upper limit of the number of tellers per day.

[0105] Since a time period is divided into two time slots, a working day has 16 working time slots calculated according to 8 working time slots. The number of tellers assigned to each time slot cannot exceed the upper limit of the number of tellers per day. Therefore, the constraint condition of formula (9) is that the sum of the number of tellers assigned to all time slots cannot exceed 16 times the upper limit of the number of tellers per day. Of course, in actual application, the multiple of N in formula (9) can be adaptively adjusted according to the number of time slots.

[0106] λ i is the number of customers arriving in each half period, μ i is the average service time of the teller in each half period. To ensure that the waiting line does not always increase, the constraint condition of the above formula (10) needs to be met.

[0107] Formula (11) is an integer constraint, because the final solution n i is the number of tellers, so it needs to satisfy the integer constraint.

[0108] t i is the average waiting time of customers in each half period, and T is the maximum waiting time of customers that the business can tolerate. Therefore, the constraint of formula (12) is that the average waiting time of customers in each half period cannot exceed the maximum waiting time of customers that the business can tolerate.

[0109] Solving the above planning problem can obtain the optimal daily teller scheduling, that is, Minimum, t is known through the queuing theory formula i Yes i function, so the global optimal solution can be found, which is the total scheduling combination formed by selecting the optimal number of people in each time period.

[0110] Assume that the optimal number of tellers in each period is but Optimal The following conditions must be met:

[0111]

[0112] After expanding the formula, it becomes:

[0113]

[0114] Can be transformed into:

[0115]

[0116] That is, in each period, we find To satisfy the above formula, the This can form the optimal scheduling of tellers.

[0117] In some embodiments of the present application, the daily upper limit of the number of tellers is determined by: determining the busiest time period based on the historical customer arrival data; determining an estimated number of customer arrivals during the daily busiest time period based on the estimated number of customer arrivals in the future time period; determining the daily upper limit of the number of tellers based on the estimated number of customer arrivals during the daily busiest time period and the average waiting time of satisfied customers, as well as a functional relationship between the number of tellers and the average waiting time of customers.

[0118] By conducting statistical analysis on historical data, we can determine the busiest time of the day. For example, the busiest time is generally between 9:30 and 10:30 a.m., when the number of incoming customers accounts for approximately 8.9% of the total number of customers for the entire day. The estimated number of customers during this time period is an important data for estimating the upper limit of the number of tellers for that day.

[0119] Assuming that the average waiting time of customers who are satisfied with the service is T, the minimum number of tellers that meets T can be selected from the functional relationship between the number of tellers and the average waiting time of customers obtained in the above embodiment. In this way, the minimum number of tellers required per day to meet the business requirements can be calculated, thereby completing the calculation of the daily upper limit N of the number of tellers for the next month.

[0120] In some embodiments of the present application, converting the integer programming problem of cloud counter teller scheduling into a QUBO model includes: defining binary decision variables; and constructing the QUBO model based on the binary decision variables and the integer programming problem of cloud counter teller scheduling.

[0121] Based on the principle of the QUBO model, when converting the integer programming problem of cloud counter teller scheduling into the QUBO model, the embodiment of the present application can first define the binary decision variable x in the cloud counter teller scheduling scenario: vm , for example, we can define vm =1 means that the node m of time slot v (the number of tellers corresponding to time slot v) is selected, x vm =0 means that the node m of time slot v is not selected, and the selected nodes can constitute the final optimal combination.

[0122] By converting the scheduling problem of cloud counter tellers into a QUBO model, it facilitates subsequent problem solving and reduces the difficulty of problem solving.

[0123] In some embodiments of the present application, constructing the QUBO model based on the binary decision variables and the integer programming problem of cloud counter teller scheduling includes: constructing the loss function of the QUBO model based on the binary decision variables and the loss function of the integer programming problem of cloud counter teller scheduling; constructing the constraints of the binary decision variables, and the constraints of the binary decision variables are configured so that one scheduling time period corresponds to one number of tellers; adjusting the loss function of the QUBO model according to the constraints of the binary decision variables to obtain the adjusted loss function of the QUBO model.

[0124] Based on the principle of the QUBO model, the loss function of the above integer programming problem can be transformed into the following QUBO model:

[0125]

[0126] Since only one teller number arrangement can be selected for a time slot, it can be considered that a node can only be assigned to one time slot, so the following constraints hold:

[0127]

[0128] Furthermore, the loss function of formula (18) can be transformed into the following form:

[0129]

[0130] Among them, S is the number of The penalty coefficient for this constraint condition can be set, for example, to be v ∑m (an m +bt(n m ))x vm An order of magnitude larger.

[0131] In some embodiments of the present application, the QUBO model is solved by using a preset quantum solution algorithm to obtain the number of cloud counter tellers in each future scheduling time period, including: the QUBO model is solved by using a quantum solution algorithm of a coherent Ising machine or a quantum annealing machine to obtain the number of cloud counter tellers in each future scheduling time period.

[0132] The QUBO model is suitable for solving coherent Ising machines (CIM) or quantum annealing machines. The solution of the physical machine takes the CIM based on the degenerate optical parametric oscillator (DOPO) as an example. This is a hybrid quantum computing system consisting of an optical part and an electrical part. The optical part includes a laser, an amplifier, a periodically polarized lithium niobate crystal (PPLN) and a fiber loop. The laser uses a femtosecond pulse fiber laser and an amplifier system. The amplified laser is first frequency-doubled using a PPLN crystal. The frequency-doubled laser is used as a pump source to synchronously pump a PPLN crystal in a fiber loop to form a degenerate optical parametric oscillation. Hundreds of oscillation pulses can exist simultaneously in the fiber loop. The electrical part includes FPGA (Field Programmable Gate Array), AD / DA (digital analog / analog digital conversion), and phase detection. The laser output from the fiber loop and the baseband laser are measured by a phase detector, and the phase of the output light can be tested. FPGA is used with high-speed AD / DA to measure and feedback control optical pulses. Unlike classical computers running on semiconductor integrated circuits, CIM uses laser pulses in optical fibers as quantum bits for calculations. In DOPO, the pump light is incident on the nonlinear optical crystal and split into two beams of light. The two beams have the same polarization direction and half the frequency of the pump light. They are in a compressed state and can be used as a quantum bit. Gradually increase the power of the pump light. When it exceeds the oscillation threshold, the generated light becomes a coherent state, and the phase of the light is divided into two states (phase 0 state and π state). At this time, the phase can be set to ±1 of the spin to solve the optimization problem.

[0133] The embodiment of the present application also provides a scheduling device 300 for a teller at a bank cloud counter, such as Figure 3As shown, a schematic diagram of the structure of a scheduling device for a teller at a bank cloud counter in an embodiment of the present application is provided. The scheduling device 300 for a teller at a bank cloud counter includes: an acquisition unit 310, a determination unit 320, a construction unit 330, a conversion unit 340, and a solution unit 350, wherein:

[0134] An acquisition unit 310 is used to acquire historical customer arrival data;

[0135] A determination unit 320, configured to determine an estimated value of the number of customer arrivals in a future time period using a customer arrival number prediction algorithm based on the historical customer arrival data;

[0136] A construction unit 330 is used to construct an integer programming problem for scheduling cloud counter tellers based on a queuing theory model according to an estimated number of customers arriving in a future time period and service data of cloud counter tellers;

[0137] A conversion unit 340, configured to convert the integer programming problem of the cloud counter teller scheduling into a QUBO model;

[0138] The solving unit 350 is used to solve the QUBO model using a preset quantum solving algorithm to obtain the number of cloud counter tellers in each future scheduling time period.

[0139] In some embodiments of the present application, the conversion unit 340 is specifically used to: define binary decision variables; and construct the QUBO model according to the binary decision variables and the integer programming problem of the cloud counter teller scheduling.

[0140] In some embodiments of the present application, the conversion unit 340 is specifically used to: construct the loss function of the QUBO model based on the binary decision variables and the loss function of the integer programming problem of the cloud counter teller scheduling; construct the constraints of the binary decision variables, and the constraints of the binary decision variables are configured so that one scheduling time period corresponds to one number of tellers; adjust the loss function of the QUBO model according to the constraints of the binary decision variables to obtain the adjusted loss function of the QUBO model.

[0141] In some embodiments of the present application, the solving unit 350 is specifically used to solve the QUBO model using a quantum solving algorithm of a coherent Ising machine or a quantum annealing machine to obtain the number of cloud counter tellers in each future scheduling time period.

[0142] In some embodiments of the present application, the customer arrival number prediction algorithm includes a time series autoregressive model, and the determination unit 320 is specifically used to: input the historical customer arrival data into the time series autoregressive model to obtain the probability distribution of the number of customer arrivals for each day in the future output by the time series autoregressive model; average the probability distribution of the number of customer arrivals for each day in the future to obtain the number of customer arrivals for each day in the future; perform statistical analysis on the historical customer arrival data to calculate the proportion of the number of customer arrivals for each time period of each day in the future to the number of customer arrivals for the entire day; calculate the estimated value of the number of customer arrivals for each time period of each day in the future based on the number of customer arrivals for each day in the future and the proportion of the number of customer arrivals for each time period of each day to the number of customer arrivals for the entire day.

[0143] In some embodiments of the present application, the construction unit 330 is specifically used to: construct a functional relationship between the number of tellers and the average waiting time of customers based on a queuing theory model according to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers; and construct an integer programming problem for scheduling cloud counter tellers based on the functional relationship between the number of tellers and the average waiting time of customers.

[0144] In some embodiments of the present application, the construction unit 330 is specifically used to: perform statistical analysis on the historical customer arrival data to determine the distribution characteristics of the number of customer incoming calls, the time interval of customer incoming calls, and the teller service time; determine the target queuing theory model based on the distribution characteristics of the number of customer incoming calls, the time interval of customer incoming calls, and the teller service time; determine the average number of customer incoming calls per second in each time period based on the estimated number of customer arrivals in each time period every day in the future; perform statistical analysis on the historical customer arrival data to calculate the mean of the teller service time; and construct a functional relationship between the number of tellers and the average waiting time of customers based on the target queuing theory model based on the average number of customer incoming calls per second and the mean of the teller service time in each time period.

[0145] It can be understood that the above-mentioned scheduling device for bank cloud counter tellers can implement the various steps of the scheduling method for bank cloud counter tellers provided in the aforementioned embodiments. The relevant explanations on the scheduling method for bank cloud counter tellers are applicable to the scheduling device for bank cloud counter tellers and will not be repeated here.

[0146] Figure 4 Schematic diagram of the structure of a device in the embodiment of the present application. Figure 4 As shown, the device includes one or more processors (or processing units), may further include one or more memories coupled to the processors, and may further include a communication module coupled to the processors.

[0147] The communication module can be used to communicate with other devices or apparatuses, such as the transmission or reception of data and / or signals. The communication module can have at least one communication module for communication. The communication module can include any interface necessary for communicating with other devices. Exemplarily, the communication module can be a transceiver, a circuit, a bus, a module, or other types of communication modules.

[0148] The processor may include, but is not limited to, at least one of the following: a general-purpose computer, a special-purpose computer, a microcontroller, a digital signal controller (DSP), or one or more of a controller-based multi-core controller architecture. The device may have multiple processors, such as application-specific integrated circuit chips, which are time-dependent and synchronized with a clock of a main processor.

[0149] The memory may include one or more non-volatile memories and one or more volatile memories. Examples of non-volatile memories include, but are not limited to, at least one of the following: read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, hard disk, compact disc (CD), digital video disc (DVD), or other magnetic storage and / or optical storage. Examples of volatile memories include, but are not limited to, at least one of the following: random access memory (RAM), or other volatile memories that do not persist during the duration of a power outage.

[0150] The computer program includes computer executable instructions executed by an associated processor. The program can be stored in ROM. The processor can perform any suitable actions and processes by loading the program into RAM.

[0151] The possible implementation of the present application can be implemented by means of a program, so that the communication device can perform any process discussed in the above embodiments. The possible implementation of the present application can also be implemented by hardware or by a combination of software and hardware.

[0152] In some embodiments, the program may be tangibly contained in a computer-readable storage medium, which may be included in the device (such as in a memory) or other storage device accessible by the device. The program may be loaded from the computer-readable storage medium to the RAM for execution. The computer-readable storage medium may include any type of tangible non-volatile memory, such as ROM, EPROM, flash memory, hard disk, CD, DVD, etc.

[0153] The present application embodiment also provides a computer-readable storage medium, on which computer instructions or program codes are stored, and when the processor runs the instructions or the program codes, the processor executes the methods and functions involved in any of the above embodiments. Computer-readable media can be any tangible medium containing or storing programs for or related to instruction execution systems, devices or equipment. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or devices, or any suitable combination thereof. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrations. More detailed examples of computer-readable storage media include electrical connections with one or more wires, magnetic media (e.g., disks, floppy disks, hard disks, tapes, magnetic storage devices), optical media (e.g., optical storage devices, DVDs), semiconductor media (e.g., solid-state hard drives), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), or any suitable combination thereof, etc.

[0154] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The embodiment of the present application also provides at least one computer program product tangibly stored on a non-temporary computer-readable storage medium. The computer program product includes one or more computer executable instructions, such as instructions included in a program module, which are executed in a device on a real or virtual processor of the target to perform the process, method and function involved in any of the above embodiments. When the computer program instruction is loaded and executed on a computer, a process or function according to an embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instruction can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instruction can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center.

[0155] The present application embodiment also proposes a computer program product, including a computer program or instruction, when the computer program or instruction is run on a computer, the computer is made to perform the process, method and function in the above-mentioned embodiment. Usually, a program module includes routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or realize specific abstract data types. In various embodiments, the functions of program modules can be combined or divided between program modules as needed. Machine executable instructions for program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0156] In general, various embodiments of the present application may be implemented in hardware or dedicated circuits, software, logic, or any combination thereof. Some aspects may be implemented in hardware, while other aspects may be implemented in firmware or software, which may be performed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of the present disclosure are shown and described as block diagrams, flow charts, or using some other graphical representations, it should be understood that the boxes, devices, systems, techniques, or methods described herein may be implemented as, for example, non-limiting examples, hardware, software, firmware, dedicated circuits or logic, general hardware or controllers or other computing devices, or some combination thereof.

[0157] It should be noted that although the embodiments of the present application are described above in conjunction with the accompanying drawings, the above embodiments are not independent of each other, and they can also be combined to obtain other embodiments. The division of the modes, situations, categories and embodiments in the embodiments of the present application is only for the convenience of description and should not constitute a special limitation. The features in the various modes, categories, situations and embodiments can be combined with each other in a logical manner. The various implementation methods of the present application can be combined arbitrarily to achieve different technical effects. The embodiments of the present application no longer list various combinations.

[0158] In addition, although the operation of the method of the present disclosure is described in a particular order in the accompanying drawings, this does not require or imply that these operations must be performed in this particular order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flow chart can change the order of execution. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution. It should also be noted that the features and functions of two or more devices according to the present disclosure can be embodied in one device. Conversely, the features and functions of a device described above can be further divided into being embodied by multiple devices.

[0159] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0160] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for scheduling tellers at a bank cloud counter, characterized in that: The scheduling method of the bank cloud counter teller includes: Get historical customer arrival data; Determine an estimated value of the number of customers arriving in a future time period using a customer arrival number prediction algorithm based on the historical customer arrival data; Based on the queuing theory model, an integer programming problem for cloud counter teller scheduling is constructed according to the estimated number of customers arriving in the future time period and the service data of cloud counter tellers; Convert the integer programming problem of the cloud counter teller scheduling into a QUBO model; The QUBO model is solved using a preset quantum solution algorithm to obtain the number of cloud counter tellers in each future scheduling time period.

2. The method for dispatching tellers at a bank cloud counter according to claim 1, characterized in that: The converting of the integer programming problem of the cloud counter teller scheduling into a QUBO model comprises: Define binary decision variables; The QUBO model is constructed according to the binary decision variables and the integer programming problem of the cloud counter teller scheduling.

3. The method for dispatching bank cloud counter tellers according to claim 1 or 2, characterized in that: The constructing of the QUBO model according to the binary decision variables and the integer programming problem of the cloud counter teller scheduling includes: Constructing the loss function of the QUBO model according to the binary decision variables and the loss function of the integer programming problem of the cloud counter teller scheduling; Constructing the constraint condition of the binary decision variable, wherein the constraint condition of the binary decision variable is configured such that one scheduling time period corresponds to one number of tellers; The loss function of the QUBO model is adjusted according to the constraint conditions of the binary decision variables to obtain an adjusted loss function of the QUBO model.

4. The method for dispatching bank cloud counter tellers according to any one of claims 1 to 3, characterized in that: The QUBO model is solved by using a preset quantum solution algorithm to obtain the number of cloud counter tellers in each future scheduling time period, including: The QUBO model is solved using a quantum solution algorithm of a coherent Ising machine or a quantum annealing machine to obtain the number of cloud counter tellers in each future scheduling time period.

5. The method for dispatching bank cloud counter tellers according to claim 1, characterized in that: The customer arrival number prediction algorithm includes a time series autoregressive model, and the method of determining the estimated value of the customer arrival number in a future time period by using the customer arrival number prediction algorithm based on the historical customer arrival data includes: Inputting the historical customer arrival data into the time series autoregressive model to obtain a probability distribution of the number of customer arrivals per day in the future output by the time series autoregressive model; Average the probability distribution of the number of customers arriving each day in the future to get the number of customers arriving each day in the future; Perform statistical analysis on the historical customer arrival data to calculate the ratio of the number of customer arrivals in each time period of each day in the future to the total number of customer arrivals for the whole day; Based on the number of customer arrivals each day in the future and the proportion of the number of customer arrivals in each time period of each day to the total number of customer arrivals for the day, calculate the estimated number of customer arrivals in each time period of each day in the future.

6. The method for dispatching tellers at a bank cloud counter according to claim 1, characterized in that: The integer programming problem of cloud counter teller scheduling based on the queuing theory model and the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers includes: According to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers, the functional relationship between the number of tellers and the average waiting time of customers is constructed based on the queuing theory model; An integer programming problem for cloud counter teller scheduling is constructed based on the functional relationship between the number of tellers and the average waiting time of customers.

7. The method for dispatching tellers at a bank cloud counter according to claim 6, characterized in that: The functional relationship between the number of tellers and the average waiting time of customers is constructed based on the queue theory model according to the estimated number of customers arriving in the future time period and the service data of the cloud counter tellers, including: Performing statistical analysis on the historical customer arrival data to determine the distribution characteristics of the number of customer incoming calls, the time interval between customer incoming calls, and the teller service time; Determine the target queuing theory model based on the distribution characteristics of the number of customer calls, the time interval between customer calls, and the teller service time; Determine the average number of incoming customer calls per second in each period based on the estimated number of customer arrivals in each period of the future day; Performing statistical analysis on the historical customer arrival data to calculate the mean of teller service time; According to the average number of incoming customer calls per second and the mean of teller service time in each time period, the functional relationship between the number of tellers and the average waiting time of customers is constructed based on the target queuing theory model.

8. A scheduling device for bank cloud counter tellers, characterized in that: The scheduling device for the bank cloud counter teller includes: An acquisition unit, used to acquire historical customer arrival data; A determination unit, configured to determine an estimated value of the number of customers arriving in a future time period based on the historical customer arrival data and using a customer arrival number prediction algorithm; A construction unit, for constructing an integer programming problem for scheduling cloud counter tellers based on a queuing theory model and according to an estimated number of customers arriving in a future time period and service data of cloud counter tellers; A conversion unit, used for converting the integer programming problem of the cloud counter teller scheduling into a QUBO model; The solving unit is used to solve the QUBO model by using a preset quantum solving algorithm to obtain the number of cloud counter tellers in each future scheduling time period.

9. A device comprising: processor; And a memory arranged to store computer executable instructions, which, when executed, cause the processor to execute the bank cloud counter teller scheduling method described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instructions are executed by a processor, the method for scheduling bank cloud counter tellers as described in any one of claims 1 to 7 is implemented.

Images