CN113435674A

CN113435674A - Customer service resource information processing and model generating method and device

Info

Publication number: CN113435674A
Application number: CN202010209590.4A
Authority: CN
Inventors: 王骏龙
Original assignee: Alibaba Group Holding Ltd
Current assignee: Alibaba Group Holding Ltd
Priority date: 2020-03-23
Filing date: 2020-03-23
Publication date: 2021-09-24

Abstract

The embodiment of the application discloses a method and a device for scheduling customer service resources and generating a model, wherein the method comprises the following steps: periodically sending a scheduling request to a proxy server, wherein the scheduling request carries state information of a current customer service environment, the proxy server stores a policy control model, and the policy control model is used for inputting the received state information of the current customer service environment into the policy control model and outputting corresponding action information; and scheduling the current customer service resource according to the action information returned by the proxy server. By the aid of the method and the device, the skill set can be more efficiently scheduled.

Description

Customer service resource information processing and model generating method and device

Technical Field

The application relates to the technical field of customer service resource information processing, in particular to a customer service resource information processing and model generating method and device.

Background

In some service systems, a customer service system is usually provided for users, and the users can consult, feedback or complain about problems through the customer service system. For example, in a commodity object network selling system, there are a large number of buyer users and seller users, and normally, the buyer users can communicate with the seller users some information including goods returns, etc., and can complain to the customer service system of the system when the negotiation is not successful or the divergence occurs, etc.

In summary, customer service systems are often required to service a very large number of users, and in some large systems, may even be required to service hundreds of millions of members. This means that a great deal of customer service resources, including personnel or robots, need to be provided in the customer service system in order to respond to the service requests of the users in a timely manner. In addition, because there may be many types of problems that the user may submit, and the same customer service resource may only solve one or a few types of problems, usually, a plurality of skill groups may be established according to a specific problem type, a specific customer service resource may be divided into corresponding skill groups according to the types of problems that the specific customer service resource can solve, and the same customer service resource may be capable of solving a plurality of types of problems, and therefore, may also be located in a plurality of skill groups. In addition, there may be multiple skill sets for the same question type. Therefore, after the service request of the user is submitted, the specific customer service resource can be selected from the corresponding skill set according to the specific corresponding problem type to provide the service for the user.

In addition, in practical applications, in order to maximize the service efficiency of the customer service resources, a scheduling system may be provided. The scheduling system has the function of executing a scheduling strategy on a specific skill set according to the condition of a specific user service request and the states of customer service resources in various different skill sets, so that the waiting time of a user is reduced as much as possible, and the user experience is improved. For example, a service request of a user enters a waiting queue of skill set a, however, all the customer service resources in skill set a are in a service state, and if not scheduled, the user can only wait for the queue in skill set a to continue waiting. However, at this time, the customer service resource in skill group B may actually be for solving the problem of the user, and there is currently a free customer service resource in skill group B, so the service request of the user may be overflowed into skill group B by the scheduling system, so that the service request of the user is responded faster, and so on.

However, in the prior art, the scheduling system described above belongs to a semi-manual scheduling mode, that is, an operator is required to submit a rule in the scheduling system, and when the rule is triggered, a certain scheduling policy is executed. This practice is very demanding for the operator and also very labor intensive. For example, the total number of skill sets in a system may be hundreds, even if only 30 rules are assigned to each skill set, then tens of thousands of rules may be assigned. Secondly the operator will often adjust the rules, for example in some cases the conditions under which skill set a can spill over to skill set B may become more stringent, but the conditions under which skill set a spills over to skill set C may become less stringent, etc. If the adjustment is made once every 2 days for each overflow combination, then one operator may work a day just to adjust the rules.

It can be seen that how to more efficiently implement the scheduling of the skill set becomes a technical problem to be solved by those skilled in the art.

Disclosure of Invention

The application provides a customer service resource information processing method and a customer service resource information processing device and a customer service resource information processing model generating device, which can realize more efficient scheduling of skill sets.

The application provides the following scheme:

a customer service resource information processing method comprises the following steps:

periodically sending a request for acquiring policy information to a proxy server, wherein the request carries state information of a current customer service environment, the proxy server stores a policy control model and is used for acquiring a predicted value of a service request quantity within a future preset time length according to the received state information of the current customer service environment, inputting the predicted value and the state information of the customer service environment into the policy control model, and outputting corresponding action information by the policy control model; the action information comprises policy execution mode information;

and executing a corresponding strategy according to the action information returned by the proxy server, wherein the strategy is used for adjusting the state information of the current customer service environment so as to balance the service efficiency within the future preset time span.

A method of providing customer service resource information, comprising:

receiving a request for acquiring policy information sent by a scheduling system, wherein the request carries state information of a current customer service environment;

obtaining a service request predicted value in a future preset time length;

inputting the service request predicted value and the state information of the current customer service environment into the strategy control model, and outputting corresponding action information by the strategy control model; the action information comprises policy execution mode information;

and returning the action information to the scheduling system so that the scheduling system executes a corresponding strategy, wherein the strategy is used for adjusting the state information of the current customer service environment so as to balance the service efficiency within the future preset time span.

A method of generating a policy control model, comprising:

simulating and generating a customer service environment;

periodically acquiring state information of a simulated customer service environment, inputting the state information into a strategy control model, and adjusting the state of the simulated environment according to action information returned by the strategy control model;

and counting the service efficiency information generated by the adjusted state, and providing the service efficiency information as an incentive value to the strategy control model so as to optimize the strategy control model according to the incentive value.

the client service business system periodically acquires policy information to be executed from the associated proxy server according to the state information of the client service environment;

in the process of adjusting the state information of the customer service environment by executing the corresponding policy, if the adjustment of the customer service resource information is involved, an adjustment request is sent to a customer service resource provider client so that the customer service resource provider client can execute the corresponding adjustment operation.

a client of a customer service resource provider receives an adjustment request for adjusting customer service resource information, wherein the adjustment request is generated by a service party customer service system after acquiring policy information to be executed from a related proxy server according to state information of a customer service environment;

and executing a corresponding adjustment action according to the adjustment request, and returning an adjustment result to the service side client service system.

the client service system periodically sends a request for acquiring the policy information to the client service resource providing system, wherein the request carries service state information and skill set state information of the client service resource, so that the client service resource providing system acquires the policy information to be executed from the associated proxy server according to the received state information, and the shift list information and the service capability information of the client service resource provided to the client service system;

receiving the strategy information returned by the customer service resource providing system;

and adjusting the environmental state information of the customer service business system by executing the strategy information.

the client service resource providing system periodically receives a request for acquiring the strategy information sent by the client service system, wherein the request carries service state information and skill set state information of the client service resource;

determining the schedule information and the service capability information of the customer service resources provided for the customer service business system;

acquiring policy information to be executed from a related proxy server according to the received state information, the shift table information of the client service resource and the service capability information;

and returning the policy information to be executed to the customer service system, so that the customer service system adjusts the environmental state information of the customer service system by executing the policy information.

A customer service resource information processing apparatus comprising:

a request sending unit, configured to periodically send a request for obtaining policy information to a proxy server, where the request carries state information of a current customer service environment, and the proxy server stores a policy control model, and is configured to obtain a predicted value of a service request volume within a preset time length in the future according to the received state information of the current customer service environment, input the predicted value and the state information of the customer service environment to the policy control model, and output corresponding action information by the policy control model; the action information comprises policy execution mode information;

and the action execution unit is used for executing a corresponding strategy according to the action information returned by the proxy server, and the strategy is used for adjusting the state information of the current customer service environment so as to balance the service efficiency within the future preset time span.

An apparatus for providing customer service resource information, comprising:

a request receiving unit, configured to receive a request for obtaining policy information sent by a scheduling system, where the request carries state information of a current customer service environment;

the predicted value obtaining unit is used for obtaining a service request predicted value in a future preset time length;

the input unit is used for inputting the service request predicted value and the state information of the current customer service environment into the strategy control model, and the strategy control model outputs corresponding action information; the action information comprises policy execution mode information;

and the action information returning unit is used for returning the action information to the scheduling system so that the scheduling system can execute a corresponding strategy, and the strategy is used for adjusting the state information of the current customer service environment so as to balance the service efficiency within the future preset time span.

An apparatus for generating a policy control model, comprising:

the environment simulation unit is used for simulating and generating a customer service environment;

the scheduling simulation unit is used for periodically acquiring state information of the simulated customer service environment, inputting the state information into the strategy control model and adjusting the state of the simulated environment according to action information returned by the strategy control model;

and the reward value providing unit is used for counting the service efficiency information generated by the adjusted state and providing the service efficiency information as a reward value to the strategy control model so as to optimize the strategy control model according to the reward value.

A customer service resource information processing device is applied to a customer service business system and comprises:

a policy information acquisition unit for periodically acquiring policy information to be executed from an associated proxy server according to the state information of the customer service environment;

an adjustment request sending unit, configured to send an adjustment request to a customer service resource provider client if adjustment of customer service resource information is involved in adjusting the state information of the customer service environment by executing a corresponding policy, so that the customer service resource provider client performs a corresponding adjustment operation.

A customer service resource information processing apparatus applied to a customer service resource provider client, comprising:

an adjustment request receiving unit, configured to receive an adjustment request for adjusting client service resource information, where the adjustment request is generated by the service-side client service system after acquiring policy information to be executed from an associated proxy server according to state information of a client service environment;

and the adjusting action executing unit is used for executing the corresponding adjusting action according to the adjusting request and returning the adjusting result to the service side client service system.

a request sending unit, configured to periodically send a request for obtaining policy information to a customer service resource providing system, where the request carries service state information and skill set state information of a customer service resource, so that the customer service resource providing system obtains policy information to be executed from an associated proxy server according to the received state information, and shift list information and service capability information of the customer service resource provided to a customer service business system;

a policy information receiving unit, configured to receive policy information returned by the customer service resource providing system;

and the strategy execution unit is used for adjusting the environmental state information of the customer service business system by executing the strategy information.

A customer service resource information processing device applied to a customer service resource providing system includes:

a request receiving unit, configured to periodically receive a request for obtaining policy information sent by a customer service system, where the request carries service state information of a customer service resource and skill set state information;

the information determining unit is used for determining the shift table information and the service capability information of the client service resources provided for the client service business system;

the policy information acquisition unit is used for acquiring policy information to be executed from the associated proxy server according to the received state information, the shift list information of the client service resource and the service capability information;

and the policy information returning unit is used for returning the policy information to be executed to the customer service system so that the customer service system can adjust the environmental state information of the customer service system by executing the policy information.

According to the specific embodiments provided herein, the present application discloses the following technical effects:

through the embodiment of the application, the corresponding executable action information under the specific customer service environment state can be generated through the strategy control model, and then the corresponding action is executed according to the output information of the model, so that the adjustment of the customer service resource state can be completed, the adjustment in a semi-manual mode is avoided, the efficiency is improved, and the labor cost is reduced.

In a preferred embodiment, a policy control model may be trained in a reinforcement learning manner, and for this reason, an embodiment of the present application further provides a simulation system, where a skill set, a user, and the like may be simulated, various policies may be simulated, functions such as session management, queue management, and the like may be implemented under the time push of a simulation backbone, and may interact with an agent system, periodically provide state information of a simulated scheduling environment to the model, and perform scheduling according to actions output by the model, including executing a specified policy, and the like. In addition, the simulation system can also perform statistics on the system service performance information after the action given by the model is performed through a data statistics function, wherein the statistics includes the total amount of service requests received in a certain time period, the number of received service requests, the number of service requests with call loss, and the like. The service efficiency information can be used as an incentive value to be provided to the model again, so that the model can use the incentive value as an index for judging the effectiveness of the output action in the corresponding input environment state, and the model is optimized, thereby achieving the purpose of strengthening learning.

Of course, it is not necessary for any product to achieve all of the above-described advantages at the same time for the practice of the present application.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.

FIG. 1 is a schematic diagram of a system architecture provided by an embodiment of the present application;

FIG. 2 is a flow chart of a first method provided by an embodiment of the present application;

FIG. 3 is a schematic diagram of another system provided by embodiments of the present application;

FIG. 4 is a flow chart of a second method provided by embodiments of the present application;

FIG. 5 is a flow chart of a third method provided by embodiments of the present application;

6-1, 6-2 are schematic diagrams of another system provided by embodiments of the present application;

FIG. 7 is a flow chart of a fourth method provided by embodiments of the present application;

FIG. 8 is a flow chart of a fifth method provided by embodiments of the present application;

FIG. 9 is a flow chart of a sixth method provided by embodiments of the present application;

FIG. 10 is a flow chart of a seventh method provided by embodiments of the present application;

FIG. 11 is a schematic diagram of a first apparatus provided by an embodiment of the present application;

FIG. 12 is a schematic diagram of a second apparatus provided by an embodiment of the present application;

FIG. 13 is a schematic diagram of a third apparatus provided by an embodiment of the present application;

FIG. 14 is a schematic diagram of a fourth apparatus provided by an embodiment of the present application;

FIG. 15 is a schematic diagram of a fifth apparatus provided by an embodiment of the present application;

FIG. 16 is a schematic view of a sixth apparatus provided by an embodiment of the present application;

FIG. 17 is a schematic diagram of a seventh apparatus provided by an embodiment of the present application;

FIG. 18 is a schematic diagram of a computer system provided by an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments that can be derived from the embodiments given herein by a person of ordinary skill in the art are intended to be within the scope of the present disclosure.

In the embodiment of the present application, in order to implement more efficient customer service resource state adjustment, a specific scheduling system may be implemented in a fully automatic manner, and an Agent system (Agent) may be provided in the Agent system. Specifically, in the process of adjusting the state of the customer service resource, the scheduling system may initiate an inquiry to the agent system at a certain time interval (e.g., 4S, etc.), where the inquiry may carry state information of the current customer service environment, including information of a shift table of the customer service resource, a service state, a service capability, a skill set condition, and the like, and the agent system may first obtain service volume information that may be generated in the system in a future target time period, and then input the information and the state information of the current customer service environment as input information of the policy control model into the model, so that the model may output a currently executable action through operation. For example, whether and what policies are specifically implemented need to be implemented may balance the complex relationships between channels, tasks, customers, resources, etc. over a target period of time in the future. Therefore, the scheduling system executes the actions according to the actions returned by the agent system, and operators only need to pay attention to whether the actions are reasonable or not and do not need to consider the specific details of when the actions are executed, so that the efficiency can be improved, and the labor cost can be reduced. In addition, compared with manual judgment, the result obtained by the model calculation mode is more objective and accurate, and the dependence on the manual experience value is reduced.

The specific policy control model may be generated in a plurality of ways, for example, a simple way may be that, through a brute force search, various input state information that may appear is exhausted in advance, corresponding executable action information in various input states is given, and the corresponding relationships are stored in the model. In this way, the concrete model is equivalent to a search tool, and the various possible states are difficult to be exhausted, and various actions still need to be set by people, and subjective components are heavy.

For this reason, the embodiments of the present application also provide a more preferable implementation manner, that is, a policy control model is generated by means of reinforcement learning. For example, in one mode, the historical data may be counted, and the service quality information may be determined by indexes such as the number of users who have received calls in a certain period of time, the number of users who have been picked up normally, and the number of users who have lost calls (have not been picked up and have been dropped), and the like, corresponding to each of the service quality information when a policy is used or not used in various environmental conditions. Then, the information is used as known training sample information to train the strategy control model, so that the control strategy model can learn which action is used under various different states, and high service quality can be obtained.

In the embodiment of the present application, another reinforcement learning manner is provided, that is, as shown in fig. 1, a simulation system may be provided, and the simulation system is used to simulate an actual customer service resource scheduling Environment (Environment), during a simulation process, Environment State information may be provided to the agent system, a corresponding Action (Action) is provided by the model, after the simulation system executes the Action provided by the model, service performance information may be counted, and the service performance information may be used as a Reward value (Reward, i.e., a target of reinforcement learning) of the reinforcement learning and provided to the policy control model for reinforcement learning. Specifically, the policy control model may determine whether the previously output action is applicable to the corresponding environmental state according to the service performance information, and if not, the action may be adjusted when the same or similar environmental state is encountered subsequently, and so on. After multiple times of reinforcement learning, the most suitable action can be determined for various environmental state information, so that the training process of the model is completed.

The following describes in detail specific implementations provided in embodiments of the present application.

Example one

First, in this embodiment, from the perspective of an actual scheduling system, a method for processing customer service resource information is provided, and referring to fig. 2, the method may specifically include:

s201: periodically sending a request for acquiring policy information to a proxy server, wherein the request carries state information of a current customer service environment, the proxy server stores a policy control model and is used for acquiring a predicted value of a service request quantity within a future preset time length according to the received state information of the current customer service environment, inputting the predicted value and the state information of the customer service environment into the policy control model, and outputting corresponding action information by the policy control model; the action information comprises policy execution mode information;

the specific request period may depend on the actual service requirement, for example, in a specific embodiment, it may be 4S as one period, and so on. The specific request may mainly carry state information of the current customer service environment, and the specific state information may include: schedule information, service status information, service capability information, and skill set information of the customer service resources; for example, what the "Tuesday" currently on duty has, when shifts, the number of sessions that the "Tuesday" is servicing, the length of time each session has been servicing, the number of sessions that the "Tuesday" can concurrently service, the status of the skill set to which the "Tuesday" belongs (information on the number of customer service resources within the skill set, service queue, wait queue length, etc.), and so forth.

By means of the state information, analysis can be performed from multiple dimensions, and corresponding detailed information can be obtained. For example, from a particular customer service resource dimension, the specifically obtained information may be as shown in table 1:

TABLE 1

It should be noted that, in the first embodiment, an interactive process between an actual scheduling system and an agent system is described, and therefore, the state information of the client service environment in the specific scheduling request refers to information in an actual operating environment, rather than simulated information. In the process of model reinforcement learning, the model is trained by using simulated information. Specifically, the information in the actual operating environment may be obtained in various ways, for example, status information about specific shift list information, current service status, skill set, and the like may be directly read from system information; and the service capability of the customer service resource and the like can be read from the data center. The data center, which may also be referred to as a customer service factory, is mainly used for collecting and counting data generated by each customer service resource in a daily service process, so as to judge service capacity of each customer service resource.

The proxy server stores the strategy control model, so the information can be input into the model as input information after being received, and the model outputs corresponding action information after being calculated. The specific action information may include whether or not to execute the policy, if so, what policy needs to be executed, when to execute, etc. The specific policy may be various, including overflowing a certain service request from one skill group to another skill group, or adjusting target customer service resources from one skill group to another skill group, or adjusting the maximum number of concurrent sessions of part or all of the customer service resources in a certain skill group by +1, and the like.

During specific implementation, the request submitted by the scheduling system may also carry precondition information of the service request in the waiting queue associated with the skill group, so that the specific model may adjust the service request among different skill groups according to the precondition information. The specific precondition can be various, for example, the service request type information selected by the user associated with the service request (for example, the service request type information can be selected through a number key in a voice prompt), and/or the personalized information of the user associated with the service request. The personalized information may be related to the assignment or adjustment of skill sets, and may specifically include, for example, tag information of the user, gender information, information of a recently purchased commodity object, and the like. By distributing or adjusting the service skill set through the information, the satisfaction degree of the user on the service result can be improved.

In addition, since the information of the number of service requests in the current or future period of time can be considered when adjusting the number of concurrent sessions that can be serviced by the customer service resources in the skill set, for example, if the number of service requests is large, then it can be considered to adjust the maximum number of concurrent sessions of all the customer service resources in the skill set by +1, and so on. For this reason, in a specific implementation, the request may also carry traffic source information associated with the service request, since the traffic source may affect the concurrency amount of the service request, for example, if the traffic source is a sales promotion meeting place, the concurrency amount of the service request may be very large, and therefore, the number of concurrent sessions of the customer service resource may be adjusted by +1, and so on.

It should be noted that, in the specific implementation, when the specific model is calculated, whether to execute a certain policy, the execution time, and the like may also be determined according to a conflict situation between customer service resources, or whether the length of the waiting queue in a certain skill group reaches a critical point of the number of service requests that can be served by the skill group.

S202: and executing a corresponding strategy according to the action information returned by the proxy server, wherein the strategy is used for adjusting the state information of the current customer service environment so as to balance the service efficiency within the future preset time span.

And after receiving the specific action information, the scheduling system schedules the current customer service resource according to the action information. For example, if the number of service requests is very large in the environment at a certain time, the output result of the model may be the maximum number of concurrent sessions +1 of all customer service resources in a certain skill group, the attribute information of the customer service resources in the corresponding skill group is modified according to the action, and then the management of sessions and queues is performed on the basis of the modified attribute information.

The specific policy control model may be configured to calculate and determine whether a policy may be executed according to the predicted value of the service request volume within the future preset time period and the state information of the current customer service environment, so that service performance within the future preset time period is balanced, and if so, determine execution mode information of the policy, generate corresponding action information, and output the corresponding action information.

In a specific implementation, as described above, the policy control model may be generated in a reinforcement learning manner, wherein in the learning process, a simulation system may provide state information of a simulated customer resource scheduling environment for the policy control model, so that the policy control model outputs executable action information, the simulation system executes a corresponding action, and provides service performance information generated at each time step as a reinforcement learning reward value for the policy control model, so that the policy control model optimizes output actions in the same environment state; the service performance information includes: the number of service requests generated, the number of service requests picked up, and the number of service requests for call loss at the current time step.

Specifically, the structure of the simulation system in the embodiment of the present application may be as shown in fig. 3, where the simulation system mainly includes the following subsystems:

the service volume estimation subsystem: the method is used for counting the total amount of services generated in the system, the time distribution and the like in a historical time period (the time length can be the same as the time length of a future target time period corresponding to the specific prediction time, namely, the learning process is also used for balancing the relation of each party in the time length). The specific statistical manner may be various, and for example, may include a real-time sequence modeling method based on statistics (weighted smoothing, exponential smoothing, random walk, STL decomposition, ARIMA linear model, etc.), or a machine learning deep learning method (xgboost, lstm, etc.) that extracts and transforms time-series features, and so on.

Skill set simulation subsystem: customer service is used as a resource, the service volume can be continuously consumed, and the main attributes of the customer service resource comprise: average processing duration, standard deviation, current status (in service, in vacation, offline, acw, etc.), belonging shift (on-duty time, off-duty time), skill, etc. The information of the affiliated shift can be directly provided by a shift table system, and the average processing time can be generated by a data center such as a customer service factory. The customer service factory processes different cases according to the skill groups to which the Xiao-Er belongs, obtains the data according to the past, and directly estimates the distribution of processing duration with the standard deviation in the subsequent simulation, thereby simulating the duration of the one-time case. The skill set simulation subsystem is mainly used for managing a customer service resource list, controlling the percentage of the customer service resources in a small section, maintaining the state of the customer service resources, judging whether the current skill set has spare customer service resources, providing state indexes and the like.

The user generates the subsystem: the service quantity estimating subsystem estimates the service quantity and the time distribution information in a time, so that the service request quantity which is probably generated in each time step (for example, 1s) can be determined according to the information, and further a specific service request can be simulated according to the service quantity estimating subsystem, wherein each service request can correspond to one user. Subsequent simulator simulation backbones will read in specific service requests from the user generation subsystem, assign skill sets and place them into waiting queues in the groups, and so on. And then, parameter estimation can be carried out on the user of the service, and the core pre-estimation index is the maximum waiting time. Since the simulator itself usually needs to count the number of calls for loss to determine the quality of service, it is a very important parameter to determine whether a user will leave in the next second if a user has not been picked up during a certain period of time. Because the difference exists between the services, the characteristic characteristics of the services need to be taken into account, and therefore, the estimation can be carried out according to the long-time parameter distribution of the services. Since most cases are not call loss, the data is basically right missing data (only know that the user waits at least x seconds, but not know how many seconds to wait at most), so the conditional probability estimation can be used in the estimation. In the simulation process, whether the user in the waiting queue will be in call loss can be estimated, specifically, whether the user will leave in the next second can be predicted according to the waiting time of the user, the problem category corresponding to the specific request and the like. The prediction is mainly used to calculate the call loss amount in a certain period of time so as to determine the service efficiency of the system. In addition, because the specific scheduling strategy and the like are given by the model, the specific service efficiency also reflects whether the result given by the model is reasonable or not, and the model can be optimized according to the result.

The strategy simulation subsystem: all policies may be property classified so that different interfaces are invoked when a policy is executed. For example, the flow control and peak voice belong to the modification of flow attributes, overflow belongs to the modification of queue attributes, customer service is hard +1 and soft +1, and the ratio control of the rest belongs to the modification of customer service attributes, so that the future strategy is convenient to increase. The triggering mode can provide two types of condition triggering and model triggering, wherein the former can carry out testing and early experiments, and the latter is a model for facilitating reinforcement learning.

The simulation execution subsystem: the system is composed of a strategy list (list), a waiting queue, a service queue and a skill set simulator. The simulator can interact with the simulator main logic to generate a session, finish the session, maintain the queue state and the like. In addition, after the model in the agent system gives the action to be executed, the action can be executed according to the action, and the corresponding strategy is executed.

The simulator simulates a backbone: the simulation backbone is responsible for promoting 1 second and 1 second, informing the simulation execution subsystem that the time reaches a certain point, shunting the flow, acquiring a user waiting to be served from the user generation subsystem, and inquiring whether the policy control model needs to execute a certain policy.

The data statistics subsystem: the system is responsible for interacting with the simulation execution subsystem to obtain current service efficiency data, processing and calculating the data, providing the data to the strategy control model, and outputting the data to a foreground. The data output to the policy control model may mainly include: the service performance information per second includes the number of service requests generated together, the number of service requests received normally, the number of service requests lost, etc., and these data can be used to measure whether the last executed policy is appropriate, thereby helping the model to perform reinforcement learning.

Therefore, the simulation system provides a simulation environment for the model, and different reward values (targets in the reinforcement learning problem) can be obtained when different strategies are used under the same environment state, so that the reinforcement learning purpose is achieved. In addition, more Environment states can be changed for the model to learn through operations such as increasing the flow and reducing the number of customer service resources.

It should be noted that, in practical applications, since the customer service resources in the scheduling system may be provided by a specific customer service resource provider, in order to enable the provider to better match the requirements of the scheduling system and avoid situations such as surplus human resources recorded by the provider, the specific scheduling system may further provide the adjusted service state information and the correspondingly generated service performance information to the customer service resource provider user client during the process of scheduling the customer service resources and adjusting according to the output result of the model, so that the provider user client performs recording and dispatching of the customer service resources according to the service state information and the service performance information.

In a word, through the embodiment of the application, the corresponding executable action information under the specific scheduling environment state can be generated through the strategy control model, and then the scheduling process can be completed by executing the corresponding action according to the output information of the model, so that scheduling in a semi-manual mode is avoided, the efficiency is improved, and the labor cost is reduced.

Example two

The second embodiment corresponds to the first embodiment, and provides a method for providing customer service resource information from the perspective of an agent system, and referring to fig. 4, the method may specifically include:

s401: receiving a request for acquiring policy information sent by a scheduling system, wherein the request carries state information of a current customer service environment;

s402: obtaining a service request predicted value in a future preset time length;

due to the implementation, the scheduling usually aims to balance or optimize the service performance for a certain period of time, for example, the combined call loss rate in 30 minutes is low, and so on. Therefore, after receiving the scheduling request, the specific policy control model may first obtain a predicted value of the service request within a preset time duration in the future, and provide action information corresponding to the current service environment state of the client according to an actual prediction condition. The specific predicted value may be obtained in various ways, for example, the predicted value may be obtained by combining the historical number of service requests in the same time period, and the like. Specifically, according to the difference of date attributes such as working days, holidays and the like, the same period of the same day of each week may show the same or similar characteristics in terms of the number and distribution of service requests, so that the service requests in the corresponding period of time in the future can be predicted according to the distribution of the service requests in the same period of the same day in the history, and the like.

S403: inputting the service request predicted value and the state information of the current customer service environment into the strategy control model, and outputting corresponding action information by the strategy control model; the action information comprises policy execution mode information;

since the reinforcement learning process of the model is completed in advance, after specific input information is input to the model, the model can automatically output a corresponding result, that is, action information matched with the input environmental state information. Of course, in a specific implementation, except that a certain policy needs to be executed for some states, the policy may not need to be executed at more times, and therefore, the non-execution policy also belongs to one type of specific action information in the specifically output action information, and at this time, the scheduling system may continue to execute, and the specific policy does not need to be executed.

S404: and returning the action information to the scheduling system so that the scheduling system executes a corresponding strategy, wherein the strategy is used for adjusting the state information of the current customer service environment so as to balance the service efficiency within the future preset time span.

After the model outputs the specific action information, the agent system may provide it to the scheduling system, and then the scheduling system executes according to the action information.

EXAMPLE III

The third embodiment mainly provides a method for generating a policy control model from the perspective of the simulation system shown in fig. 3, and referring to fig. 5, the method may specifically include:

s501: simulating and generating a customer service environment;

specifically, when generating a customer service environment in a simulation manner, there may be a plurality of specific implementation manners, for example, in one of the manners, first, information of the number of service requests and time distribution information in a historical time period may be obtained; and then, generating a user object according to the simulation of the preset time step according to the service request quantity information and the time distribution information, and generating a corresponding service request. In addition, customer service resource information can be generated in a simulated mode, and the customer service resource information comprises: schedule information, service state information, capability information, and skill set information of the customer service resource. Then, under the push of the simulation backbone, the service request can be read in along with the advance of time, the service request is distributed to the corresponding skill group according to the associated question class information, and the waiting queue and the service queue associated with the skill group are updated. In this process, scheduling policy list information may be simulated and currently simulated customer service environment state information may be periodically input to the policy control model to obtain corresponding action information, and, if a specified policy needs to be used, corresponding policy content may be read from the policy list and the policy may be executed. In addition, the number of the service requests accessed in the last time step, the number of the service requests received and call loss conditions can be counted and provided to the strategy control model as an incentive value for reinforcement learning.

S502: periodically acquiring state information of a simulated customer service environment, inputting the state information into a strategy control model, and adjusting the state of the simulated environment according to action information returned by the strategy control model;

s503: and counting the service efficiency information generated by the adjusted state, and providing the service efficiency information as an incentive value to the strategy control model so as to optimize the strategy control model according to the incentive value.

Example four

In practical applications, a specific proxy server may be deployed in a variety of different ways. For example, in one manner, if a specific customer service resource can be managed by a specific customer service system (i.e., a system that needs to use the customer service resource to provide customer service to the outside, such as a customer service system of a bank or a customer service system of a certain e-commerce platform, etc.), the proxy server may be directly deployed in the customer service system. In yet another case, a particular customer service resource may be provided by a particular customer service resource provider (e.g., a particular customer service resource outsourcing company, etc.), and a shift list, service capability information, etc. regarding the particular customer service resource may need to be managed by the customer service resource provider. For example, because a customer service resource may be at the office of a customer service resource provider to provide services to a customer service business system, it may be desirable for the customer service resource provider to manage when a particular customer service resource may change its duty, whether it is time to change duty or shift duty, whether it is desirable to increase the number of concurrent sessions for the customer service resource, whether it is desirable to increase or decrease the allocation of customer service resources to a particular business system, and the like.

In view of the above, in the process of specifically adjusting the client service environment state of the business party, the business party and the resource provider may need to cooperate with each other. Therefore, in this case, as shown in fig. 6-1, a specific proxy server may be deployed on the side of the client service business system, or, as shown in fig. 6-2, may also be deployed on the side of the client service resource provider, for the former, a specific policy generation process may be completed by the business party through the associated proxy server, but only when the policy is executed, if adjustment of the information of the client service resource itself is involved, the client service resource provider may be requested to perform adjustment, for example, to prolong the on-duty time of the client service resource, and so on.

For the latter, a specific policy may be generated by the customer service resource provider, and at this time, the schedule information, the service capability information, and the like of the specific customer service resource may be stored locally in the customer service resource providing system, and need not be provided to the specific customer service business system; therefore, when the client service business system needs to acquire the policy information, the client service resource providing system can be provided with the information of the service state, the skill set state and the like of the specific client service resource, so that the client service resource providing system can acquire the policy information to be executed from the associated proxy server according to the received state information, the locally stored shift list information and the service capability information of the client service resource, and then provide the policy information to the client service business system to execute the corresponding policy.

Specifically, the fourth embodiment provides a customer service resource information processing method from the perspective of a customer service business system for the deployment manner shown in fig. 6-1, and with reference to fig. 7, the method may specifically include:

s701: the client service business system periodically acquires policy information to be executed from the associated proxy server according to the state information of the client service environment;

s702: in the process of adjusting the state information of the customer service environment by executing the corresponding policy, if the adjustment of the customer service resource information is involved, an adjustment request is sent to a customer service resource provider client so that the customer service resource provider client can execute the corresponding adjustment operation.

The proxy server stores a policy control model, and is configured to obtain a predicted value of a service request volume within a preset time length in the future according to received state information of a customer service environment, input the predicted value and the state information of the customer service environment to the policy control model, and output, by the policy control model, the policy information to be executed.

Specifically, the adjusting of the customer service resource information may include: adjustment of the time to shift the customer service resources, or adjustment of the number of customer service resources. For example, in the event that the number of service requests for a particular customer service business system is significant, the shift of the customer service resource may be delayed or the dispatch of the customer service resource for that customer service business system may be increased. Or, in the policy generation process in the embodiment of the present application, the predicted value of the service request amount in a future period of time is referred to, so that in the case that the resource of the customer service resource provider system is insufficient, the temporary customer service resource may be additionally recorded, and the like. In short, in this way, the customer service resource provider can perform the registration of the customer service resource and the dispatch of the customer resource to the specific worker according to the specifically generated strategy, and the strategy is generated by the model calculation, so that the blindness and subjectivity in the aspects of employee registration and dispatch can be reduced.

EXAMPLE five

The fifth embodiment is corresponding to the fourth embodiment, and provides a method for processing customer service resource information from the perspective of a customer service resource provider client, and with reference to fig. 8, the method may specifically include:

s801: a client of a customer service resource provider receives an adjustment request for adjusting customer service resource information, wherein the adjustment request is generated by a service party customer service system after acquiring policy information to be executed from a related proxy server according to state information of a customer service environment;

s802: and executing a corresponding adjustment action according to the adjustment request, and returning an adjustment result to the service side client service system.

EXAMPLE six

The sixth embodiment is directed to the case shown in fig. 6-2, and provides a method for processing customer service resource information from the perspective of a customer service business system, and with reference to fig. 9, the method may specifically include:

s901: the client service system periodically sends a request for acquiring the policy information to the client service resource providing system, wherein the request carries service state information and skill set state information of the client service resource, so that the client service resource providing system acquires the policy information to be executed from the associated proxy server according to the received state information, and the shift list information and the service capability information of the client service resource provided to the client service system;

s902: receiving the strategy information returned by the customer service resource providing system;

s903: and adjusting the environmental state information of the customer service business system by executing the strategy information.

EXAMPLE seven

The seventh embodiment is corresponding to the sixth embodiment, and provides a method for processing customer service resource information from the perspective of a customer service resource provider client, and referring to fig. 10, the method may specifically include:

s1001: the client service resource providing system periodically receives a request for acquiring the strategy information sent by the client service system, wherein the request carries service state information and skill set state information of the client service resource;

s1002: determining the schedule information and the service capability information of the customer service resources provided for the customer service business system;

s1003: acquiring policy information to be executed from a related proxy server according to the received state information, the shift table information of the client service resource and the service capability information;

s1004: and returning the policy information to be executed to the customer service system, so that the customer service system adjusts the environmental state information of the customer service system by executing the policy information.

For the parts of the second to seventh embodiments that are not described in detail, reference may be made to the description of the first embodiment, which is not described herein again.

Corresponding to the first embodiment, an embodiment of the present application further provides a client service resource scheduling apparatus, and referring to fig. 11, the apparatus may specifically include:

a request sending unit 1101, configured to periodically send a request for obtaining policy information to a proxy server, where the request carries state information of a current customer service environment, and the proxy server stores a policy control model, and is configured to obtain a predicted value of a service request amount within a preset time length in the future according to the received state information of the current customer service environment, input the predicted value and the state information of the customer service environment to the policy control model, and output corresponding action information by the policy control model; the action information comprises policy execution mode information;

an action executing unit 1102, configured to execute a corresponding policy according to the action information returned by the proxy server, where the policy is used to adjust the state information of the current customer service environment, so as to balance service performance within the preset time period in the future.

Wherein the state information of the current customer service environment includes: schedule information, service status information, service capability information, and skill set status information of the customer service resources;

wherein the skill set status information comprises: service request type information corresponding to the skill set, quantity information of customer service resources, waiting queue information and service queue information;

the action execution unit may specifically include: customer service requests that have entered a skill set waiting queue are tuned to another skill set that can handle the same type of service request.

The request also carries precondition information of the service request in the waiting queue associated with the skill group, so that the service request can be adjusted among different skill groups according to the precondition information.

Specifically, the precondition information may include: information of a service request type selected by a user associated with the service request, and/or personalized information of the user associated with the service request, the personalized information being related to the assignment or adjustment of the skill set.

In addition, the service capability information may include at least one service request type information that the customer service resource can service; the skill set status information includes: service request type information corresponding to the skill set;

in this case, the action execution unit may specifically be configured to:

target customer service resources are adjusted from one skill set to another.

In addition, the service capability information comprises information of the number of concurrent sessions which can be served by the customer service resource;

the action execution unit may specifically be configured to:

the number of concurrent sessions that can be serviced by some or all of the customer service resources in the target skill set is adjusted.

Specifically, the request may also carry traffic source information associated with the service request, so as to adjust the number of concurrent sessions of the customer service resource according to the amount of concurrent service requests associated with the traffic source.

In a specific implementation, the action information may specifically include: whether the strategy needs to be executed, and when the strategy needs to be executed, the strategy identification information and the execution time information which need to be executed.

The policy control model is used for calculating and determining whether a policy can be executed according to the predicted value of the service request volume in the future preset time span and the state information of the current customer service environment, so that the service efficiency in the future preset time span is balanced, and if so, determining the execution mode information of the policy, generating corresponding action information and outputting the action information.

In a specific implementation manner, the policy control model may be generated in a reinforcement learning manner, wherein in a learning process, a simulation system provides state information of a simulated customer resource scheduling environment for the policy control model so that the policy control model outputs executable action information, the simulation system executes a corresponding action, and provides service performance information generated at each time step as a reinforcement learning reward value for the policy control model so that the policy control model optimizes output actions in the same environment state; the service performance information includes: the number of service requests generated, the number of service requests picked up, and the number of service requests for call loss at the current time step.

In addition, when specifically implemented, the apparatus may further include:

and the information providing unit is used for providing the adjusted service state information and the correspondingly generated service efficiency information to the client of the customer service resource provider user so that the client of the provider user can record and dispatch the customer service resources according to the service state information and the service efficiency information.

Corresponding to the second embodiment, an embodiment of the present application further provides a device for providing customer service resource information, referring to fig. 12, where the device may specifically include:

a request receiving unit 1201, configured to receive a request for obtaining policy information sent by a scheduling system, where the request carries state information of a current customer service environment;

a predicted value obtaining unit 1202, configured to obtain a predicted value of a service request within a preset time length in the future;

an input unit 1203, configured to input the service request prediction value and the state information of the current customer service environment into the policy control model, and output corresponding action information by the policy control model; the action information comprises policy execution mode information;

an action information returning unit 1204, configured to return the action information to the scheduling system, so that the scheduling system executes a corresponding policy, where the policy is used to adjust the state information of the current customer service environment, so as to balance service performance within the future preset time period.

Corresponding to the three phases of the embodiment, the embodiment of the present application further provides an apparatus for generating a policy control model, referring to fig. 13, where the apparatus may specifically include:

an environment simulation unit 1301, configured to simulate and generate a customer service environment;

a scheduling simulation unit 1302, configured to periodically obtain state information of the simulated customer service environment, input the state information into the policy control model, and adjust the state of the simulated environment according to the action information returned by the policy control model;

and an incentive value providing unit 1303, configured to count service performance information generated by the adjusted state, and provide the service performance information as an incentive value to the policy control model, so as to optimize the policy control model according to the incentive value.

Specifically, the environment simulation unit may specifically include:

the service request prediction subunit is used for acquiring service request quantity information and time distribution information in a historical time period;

the user generation subunit is used for generating a user object according to the service request quantity information and the time distribution information in a simulation mode and generating a corresponding service request;

the customer service resource simulation subunit is configured to generate customer service resource information in a simulation manner, where the customer service resource information includes: schedule information, service state information, capability information and skill group information of the customer service resources;

the simulation execution subunit is used for reading the service request along with the advance of time, distributing the service request to a corresponding skill group according to the associated question class information, and updating a waiting queue and a service queue associated with the skill group;

the scheduling simulation unit may be specifically configured to simulate scheduling policy list information, periodically input state information of a currently simulated customer service environment to the policy control model to obtain corresponding action information, and if a specified policy needs to be executed, read and execute corresponding policy content from the policy list;

the reward value providing unit may be specifically configured to count the number of service requests accessed in the previous time step, the number of service requests received, and call loss, and provide the count as a reward value to the policy control model for reinforcement learning.

In addition, the present application also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor implements the steps of the method described in any of the foregoing method embodiments.

Corresponding to the fourth embodiment, an embodiment of the present application further provides an apparatus for providing customer service resource information, referring to fig. 14, where the apparatus is applied to a customer service business system, and includes:

a policy information obtaining unit 1401 for periodically obtaining policy information to be executed from an associated proxy server according to the state information of the customer service environment;

an adjustment request sending unit 1402, configured to, in the process of adjusting the state information of the customer service environment by executing the corresponding policy, send an adjustment request to a customer service resource provider client if adjustment of customer service resource information is involved, so that the customer service resource provider client performs a corresponding adjustment operation.

The adjusting of the customer service resource information comprises: adjustment of the time to shift the customer service resources, or adjustment of the number of customer service resources.

Corresponding to the fifth embodiment, the present application further provides an apparatus for providing customer service resource information, referring to fig. 15, where the apparatus is applied to a customer service resource provider client, and includes:

an adjustment request receiving unit 1501, configured to receive an adjustment request for adjusting client service resource information, which is provided by a client service system, where the adjustment request is generated by the service-side client service system acquiring policy information to be executed from an associated proxy server according to state information of a client service environment;

an adjusting action executing unit 1502, configured to execute a corresponding adjusting action according to the adjusting request, and return an adjusting result to the service-side customer service system.

Corresponding to the sixth embodiment, an embodiment of the present application further provides an apparatus for providing customer service resource information, referring to fig. 16, where the apparatus is applied to a customer service business system, and includes:

a request sending unit 1601, configured to periodically send a request for obtaining policy information to a client service resource providing system, where the request carries service state information and skill set state information of a client service resource, so that the client service resource providing system obtains policy information to be executed from an associated proxy server according to the received state information, and shift list information and service capability information of the client service resource provided to a client service system;

a policy information receiving unit 1602, configured to receive policy information returned by the customer service resource providing system;

a policy executing unit 1603, configured to adjust the environmental status information of the customer service business system by executing the policy information.

Corresponding to the seventh embodiment, an embodiment of the present application further provides an apparatus for providing customer service resource information, referring to fig. 17, the apparatus is applied to a customer service resource providing system, and includes:

a request receiving unit 1701, configured to periodically receive a request for obtaining policy information sent by a customer service system, where the request carries service state information of a customer service resource and skill set state information;

an information determining unit 1702, configured to determine shift table information and service capability information of the customer service resource provided to the customer service business system;

a policy information obtaining unit 1703, configured to obtain policy information to be executed from an associated proxy server according to the received status information, and the shift table information and the service capability information of the client service resource;

a policy information returning unit 1704, configured to return the policy information to be executed to the customer service business system, so that the customer service business system adjusts the environmental status information of the customer service business system by executing the policy information.

And a computer system, comprising:

one or more processors; and

a memory associated with the one or more processors for storing program instructions that, when read and executed by the one or more processors, perform the steps of the method described in any of the preceding method embodiments.

FIG. 18 illustrates an architecture of a computer system that may include, among other things, a processor 1810, a video display adapter 1811, a disk drive 1812, an input/output interface 1813, a network interface 1814, and memory 1820. The processor 1810, video display adapter 1811, disk drive 1812, input/output interface 1813, network interface 1814, and memory 1820 can be communicatively coupled via a communication bus 1830.

The processor 1810 may be implemented by a general CPU (Central Processing Unit), a microprocessor, an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits, and is configured to execute related programs to implement the technical solution provided by the present Application.

The Memory 1820 may be implemented in the form of a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, or the like. The memory 1820 may store an operating system 1821 for controlling operation of the electronic device 1800, and a Basic Input Output System (BIOS) for controlling low-level operation of the electronic device 1800. In addition, a web browser 1823, a data storage management system 1824, a customer service resource scheduling system 1825, and the like may also be stored. The customer service resource scheduling system 1825 may be an application program that implements the operations of the foregoing steps in this embodiment of the application. In summary, when the technical solution provided by the present application is implemented by software or firmware, the relevant program code is stored in the memory 1820 and invoked for execution by the processor 1810.

The input/output interface 1813 is used to connect input/output modules for inputting and outputting information. The i/o module may be configured as a component in a device (not shown) or may be external to the device to provide a corresponding function. The input devices may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output devices may include a display, a speaker, a vibrator, an indicator light, etc.

The network interface 1814 is used to connect a communication module (not shown in the figure) to implement communication interaction between the present device and other devices. The communication module can realize communication in a wired mode (such as USB, network cable and the like) and also can realize communication in a wireless mode (such as mobile network, WIFI, Bluetooth and the like).

The bus 1830 includes a path to transfer information between various components of the device, such as the processor 1810, the video display adapter 1811, the disk drive 1812, the input/output interface 1813, the network interface 1814, and the memory 1820.

It should be noted that although the above-described devices only illustrate the processor 1810, the video display adapter 1811, the disk drive 1812, the input/output interface 1813, the network interface 1814, the memory 1820, the bus 1830, etc., in particular implementations, the device may also include other components necessary to achieve proper operation. Furthermore, it will be understood by those skilled in the art that the apparatus described above may also include only the components necessary to implement the solution of the present application, and not necessarily all of the components shown in the figures.

From the above description of the embodiments, it is clear to those skilled in the art that the present application can be implemented by software plus necessary general hardware platform. Based on such understanding, the technical solutions of the present application may be essentially or partially implemented in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments of the present application.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, the system or system embodiments are substantially similar to the method embodiments and therefore are described in a relatively simple manner, and reference may be made to some of the descriptions of the method embodiments for related points. The above-described system and system embodiments are only illustrative, wherein the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.

The method and the device for processing the customer service resource information and generating the model provided by the application are introduced in detail, specific examples are applied in the method to explain the principle and the implementation mode of the application, and the description of the embodiments is only used for helping to understand the method and the core idea of the application; meanwhile, for a person skilled in the art, according to the idea of the present application, the specific embodiments and the application range may be changed. In view of the above, the description should not be taken as limiting the application.

Claims

1. A customer service resource information processing method is characterized by comprising the following steps:

2. The method of claim 1,

the state information of the current customer service environment includes: shift list information, service status information, service capability information, and skill set status information for the customer service resource.

3. The method of claim 2,

the skill set status information includes: service request type information corresponding to the skill set, quantity information of customer service resources, waiting queue information and service queue information;

the adjusting the state information of the current customer service environment includes:

customer service requests that have entered a skill set waiting queue are tuned to another skill set that can handle the same type of service request.

4. The method of claim 3,

5. The method of claim 4,

the precondition information includes: information of a service request type selected by a user associated with the service request, and/or personalized information of the user associated with the service request, the personalized information being related to the assignment or adjustment of the skill set.

6. The method of claim 2,

the service capability information comprises at least one service request type information which can be serviced by the customer service resource;

the skill set status information includes: service request type information corresponding to the skill set;

target customer service resources are adjusted from one skill set to another.

7. The method of claim 2,

the service capability information comprises the information of the number of concurrent sessions which can be served by the customer service resources;

8. The method of claim 7,

the request also carries traffic source information associated with the service request, so as to adjust the number of concurrent sessions of the customer service resource according to the amount of concurrent service requests associated with the traffic source.

9. The method of claim 1,

the action information includes: whether the strategy needs to be executed, and when the strategy needs to be executed, the strategy identification information and the execution time information which need to be executed.

10. The method of claim 1,

the strategy control model is used for calculating and determining whether a certain strategy can be executed according to the predicted value of the service request volume in the future preset time span and the state information of the current customer service environment, so that the service efficiency in the future preset time span is balanced, and if the strategy can be executed, the strategy execution mode information is determined, corresponding action information is generated, and then the action information is output.

11. The method of claim 1,

the strategy control model is generated in a reinforcement learning mode, wherein in the learning process, a simulation system provides state information of a simulated customer service resource scheduling environment for the strategy control model so that the strategy control model can output executable action information, the simulation system executes corresponding actions, and service efficiency information generated in each time step is provided for the strategy control model as a reinforcement learning reward value so that the strategy control model optimizes output actions in the same environment state; the service performance information includes: the number of service requests generated, the number of service requests picked up, and the number of service requests for call loss at the current time step.

12. The method of claim 11, further comprising:

and providing the adjusted service state information and the correspondingly generated service efficiency information to a client of a user provider of the customer service resource, so that the client of the user provider can record and dispatch the customer service resource according to the service state information and the service efficiency information.

13. A method for providing customer service resource information, comprising:

obtaining a service request predicted value in a future preset time length;

14. A method of generating a policy control model, comprising:

simulating and generating a customer service environment;

15. The method of claim 14,

the simulation generates a customer service environment comprising:

acquiring service request quantity information and time distribution information in a historical time period;

generating a user object according to the service request quantity information and the time distribution information and a preset time step simulation, and generating a corresponding service request;

generating customer service resource information in a simulation mode, wherein the customer service resource information comprises: schedule information, service state information, capability information and skill set information of the customer service resources;

reading the service request along with the advance of time, distributing the service request to a corresponding skill group according to the associated question class information, and updating a waiting queue and a service queue associated with the skill group;

the adjusting the state of the simulation environment comprises:

simulating scheduling policy list information, periodically inputting state information of a currently simulated customer service environment to a policy control model to obtain corresponding action information, and if a specified policy needs to be executed, reading and executing corresponding policy content from the policy list;

the counting of the service performance information generated by the adjusted state and providing the service performance information as a reward value to the policy control model includes:

and counting the number of the accessed service requests, the number of the accessed service requests and call loss conditions in the last time step, and providing the counted result as an incentive value to the strategy control model for reinforcement learning.

16. A customer service resource information processing method is characterized by comprising the following steps:

17. The method of claim 16,

the proxy server stores a strategy control model, and is used for acquiring a predicted value of service request quantity within a future preset time length according to received state information of a customer service environment, inputting the predicted value and the state information of the customer service environment into the strategy control model, and outputting the strategy information to be executed by the strategy control model.

18. The method of claim 16,

19. A customer service resource information processing method is characterized by comprising the following steps:

20. The method of claim 19,

21. A customer service resource information processing method is characterized by comprising the following steps:

22. The method of claim 21,

23. A customer service resource information processing method is characterized by comprising the following steps:

24. The method of claim 23,

25. A customer service resource information processing apparatus, comprising:

26. An apparatus for providing customer service resource information, comprising:

27. An apparatus for generating a policy control model, comprising:

28. A customer service resource information processing device is applied to a customer service business system and comprises the following components:

29. A customer service resource information processing apparatus applied to a customer service resource provider client, comprising:

30. A customer service resource information processing device is applied to a customer service business system and comprises the following components:

31. A customer service resource information processing apparatus applied to a customer service resource providing system, comprising:

32. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 24.

33. A computer system, comprising:

one or more processors; and

a memory associated with the one or more processors for storing program instructions that, when read and executed by the one or more processors, perform the steps of the method of any of claims 1 to 24.