Real-time Multiple Agent Participation Decision-making System

Abstract

To provide a real-time plural agent participation determination system that supports an autonomous determination system.SOLUTION: Based upon agent / client pairs in candidate pair paths which are an urgent agent / client pair or risky agent / client pair, an initial path value of each candidate pair path is decreased, and a final path value for the candidate pair path is derived. Based upon the final path value for the candidate pair path, the best path is determined. Based upon the best path, at least one participation determination is derived, and transmitted to an agent in the best path.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present disclosure generally relates to autonomous decision-making systems, and more specifically, to methods and related apparatuses for supporting autonomous decision-making systems.

Background Art

[0002] A multiple agent engagement decision system (MAEDS) provides timely decisions for its autonomous agents to engage with high-value clients. For this problem to occur, there must be multiple clients and multiple autonomous agents, so there are few existing MAEDSs, but in today's business, multiple client engagements by multiple agents are required. As the collective use of agents without managerial intervention becomes more prevalent, MAEDSs that adjust client engagement have become essential to ensure service success and retain important clients.

[0003] MAEDSs face several challenges in making decisions regarding multiple agent engagements. These challenges include the difficulty of making decisions regarding multiple agent engagements, the difficulty of making multiple engagement decisions that are adjusted in a timely manner, the difficulty of managing time and resources in automatically adjusting to new change drivers as the engagement evolves, and the difficulty of avoiding hasty decisions that result in waste of agent engagement capabilities and reduction of engagement options.

Summary of the Invention

[0004] Existing MAEDs are few. The reason is that MAEDs are scarce because there must be multiple clients and multiple autonomous agents, and in today's business, multiple client engagements by multiple agents are required. As the collective use of agents without administrative intervention becomes more prevalent, MAEDs that adjust client engagement are becoming essential to succeed in services and not lose important clients.

[0005] According to some embodiments of the concepts of the present invention, a method is provided for generating a set of adjusted participation parameters by a processor in a multi-agent participation decision system. The method includes determining agent / client pairs in at least one priority level based on the number of clients and the number of agents, each agent / client pair having an agent, a client, and a client rank of the client. The method further includes determining a matrix value for each agent / client pair based on the client rank of the client of the agent / client pair. The method further includes determining a participation ability parameter for each agent / client pair. The method further includes setting the matrix value to a zero value for each discarded agent / ranked client value, where the discarded agent / client pair has a participation ability margin less than a predetermined threshold. The method further includes evaluating a plurality of participation options, each participation option being a pair path from the agent / client pair with the highest client rank to the agent / client pair with the lowest client rank, each pair path having an agent / client pair for each client in at least one priority level. The method further includes calculating an initial path value for each pair path. The method further includes determining a candidate pair path by determining the pair path having the highest initial path value, where the pair path having the highest initial path value is the candidate pair path. The method further includes reducing the initial path value of each candidate pair path to derive a final path value for the candidate pair path, based on the agent / client pair in the candidate pair path that is the critical agent / client pair and based on the agent / client pair in the candidate pair path that is the risky agent / client pair. The method further includes determining the optimal path based on the final path value for each candidate pair path. The method further includes deriving at least one participation decision based on the optimal path.The method further includes sending at least one engagement decision towards an agent within an optimal path.

[0006] One advantage achievable with the concepts of the present invention is that MAEDS is a real-time system for autonomous agents and provides a set of client engagement decisions that are timely adjusted for multiple autonomous agents to serve high-value clients in a dynamically changing customer environment. Timely adjusted client engagement decisions are advantageous in preserving agent engagement capabilities, engaging with the same number of high-value clients as the number of agents (ensuring one-to-one service provision to the most valuable clients), and being flexible to changes in the client's environment. The dynamics of the client environment can include the progression of client arrival situations, the recognition of client value, engagement opportunities, management directives, agent resources, interference with engagement due to past service decisions, changes in the location and status of agents and clients, and the urgency of engagement. This improves the operation of autonomous agents because using MAEDS significantly reduces premature decisions that waste agent engagement capabilities and reduce engagement options and, in some aspects, can eliminate them.

[0007] Another advantage achievable is that MAEDS can achieve orderly decision-making without performing administrative interventions using aggregated client evaluation and recognition data, as compared to current agent engagement approaches that typically rely on only pre-planned engagement scenarios and where one agent engages with one client at a time.

[0008] In some embodiments of the concepts of the present invention, determining agent / client pairs in at least one priority level based on the number of clients and the number of agents includes dynamically obtaining from a status processor the number of clients, the number of agents, and for each client, the client rank of the client; determining the client rank of each client based on the number of clients; determining at least one priority level based on the number of agents; and grouping some agent / client pairs into at least one priority level according to the client ranks of the clients of the agent / client pairs.

[0009] In some embodiments, in response to the engagement ability parameter of an agent / client pair being within the range of a first threshold value and a second threshold value, determining that the agent / client pair is an emergency agent / client pair; and in response to the engagement ability parameter of the agent / client pair being less than a risk threshold value, determining that the agent / client pair is a risky agent / client pair, thereby determining an agent / client pair that is an emergency agent / client pair and an agent / client pair that is a risky agent / client pair.

[0010] In various embodiments of the concepts of the present invention, a value matrix of at least one priority level is derived based on the number of agents and the number of clients in at least one priority level, and the value matrix has a number of rows that is less than or equal to the number of agents for which a determination is required and less than or equal to the number of clients.

[0011] In some embodiments, deriving the value matrix includes adding lower-ranked rows to the value matrix for clients to enable more than one agent to engage with a client.

[0012] In some embodiments, each participation option is a pair path from the agent / client pair in the top row of the value matrix to the agent / client pairs in each intermediate row of the value matrix and to the agent / client pair in the bottom row of the value matrix, and each agent and client of each agent / client in the pair path is different from the other agents and clients of the agent / client pair of the pair path.

[0013] In various embodiments, the matrix value of each agent / client pair in the bottom row is set to a value of 1. The matrix value of each agent / client pair having a successively higher client rank is set to a value successively higher by a power of N-1 2. N is the number of rows in which the agent / client pairs are arranged within the value matrix, and the matrix value of the bottom row is 1.

[0014] In some embodiments of the concepts of the present invention, reducing the initial path value of a candidate pair path based on the agent / client pairs in a high-value pair path that are emergency agent / client pairs and based on the agent / client pairs in a high-value pair path that are risky agent / client pairs includes reducing the initial path value of the candidate pair path by only the matrix value of the emergency agent / client for each agent / client pair in the candidate pair path that is an emergency agent / client pair, and reducing the initial path value of the candidate pair path by only the risk value based on the participation ability parameter of the risky agent / client pair for each agent / client pair in the candidate path that is a risky agent / client pair.

[0015] In some embodiments, deriving at least one engagement decision based on the best path involves determining a first action for an agent of an emergency agent / client pair to perform in response to the best path having an emergency agent / pair and the engagement ability of the emergency agent / pair being within a decision threshold or the start time being within a time threshold, and determining a second action for an agent of an agent / client pair in the best path that is not an emergency agent / client pair in response to the agent of the agent / client pair in the best path, where the second action for the agent of the agent / client pair in the best path is to stay on course. Sending at least one engagement decision to the agent in the best path involves sending the first action to the agent of the emergency agent / client pair in response to determining the first action, and sending the second action of staying on course to the agent of the agent / client pair in the best path that is not an emergency agent / client pair.

[0016] Examples of apparatus and computer program products of the concepts of the present invention incorporate any of the above examples and permutations of the above examples of the concepts of the present invention.

[0017] The accompanying drawings, which provide a further understanding of the disclosure and are incorporated in and constitute a part of this application, illustrate specific non-limiting examples of the concepts of the present invention.

Brief Description of the Drawings

[0018]

Figure 1

Figure 2

Figure 3

Figure 4a

Figure 4b

Figure 4c

Figure 4d

Figure 4e

Figure 4f

Figure 5a

Figure 5b

Figure 6

Figure 7A

Figure 7B

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Mode for Carrying Out the Invention

[0019] Hereinafter, the concept of the present invention will be described in more detail with reference to the accompanying drawings. In the accompanying drawings, embodiments of the concept of the present invention are shown. However, the concept of the present invention may be embodied in various forms and should not be construed as being limited to the embodiments presented herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and the scope of the concept of the present invention will be fully conveyed to those skilled in the art. It should also be noted that these embodiments are not mutually exclusive. It may be implicitly assumed that the components of one embodiment are present in and used in another embodiment.

[0020] The following description presents various embodiments of the disclosed subject matter. These embodiments are presented as illustrative examples and should not be construed as limiting the scope of the disclosed subject matter. For example, certain details of the described embodiments may be modified, excluded, or extended without departing from the scope of the described subject matter.

[0021] As described above, MAEDS faces numerous challenges in decision-making involving multiple agents. These numerous challenges include the difficulty of making decisions involving multiple agents, the difficulty of making multiple, timely and adjusted decisions, the difficulty of managing time and resources in automatically adjusting to new change drivers as the involvement progresses, and the difficulty of avoiding hasty decisions that can lead to waste of agent involvement capabilities and reduction of involvement options.

[0022] Regarding the difficulty of making decisions involving multiple agents, these difficulties may include sorting out overlapping areas of client involvement capabilities, maintaining a balance between the ability to engage with the highest-value clients and the limitations of agent resources, and mobilizing all agents to provide one-on-one services to the same number of high-value customers as there are agents. Many businesses focus on one-agent and one-client engagement at a time and thus do not have the above-mentioned issues.

[0023] Regarding the difficulty of making multiple, timely and adjusted decisions, the agent's perception of the client's value changes over time, and when the information is incomplete, it becomes difficult to make timely decisions appropriately. Value fluctuations typically result from changes in the customer environment, the accuracy of recognition of such changes, and the accuracy of the client's value determination. In some cases, the customer environment may include false information regarding client value, client groups with sub-clients having different values, or service obstacles associated with the development of the involvement. When making engagement decisions alternately one by one in response to the development of customer needs or changes in location, it becomes difficult to make adjusted decisions. Past decisions may prevent future engagement options.

[0024] Regarding the difficulty of managing time and resources in automatically adjusting to new change drivers as the engagement evolves, change drivers can include the arrival of new clients, unexpected obstacles and client activities, and the unexpected health and resource capabilities of agents. Delays in managerial replanning or shortages of resources limit the adaptability of the engagement to the emergence of new change drivers. For systems with pre-planned engagement plans, attempts at replanning can take days and delay ongoing events.

[0025] Regarding the difficulty of avoiding hasty decisions that can lead to waste of agent engagement capabilities and a reduction in engagement options, such difficulties can result in poor services. The presence of multiple agents makes more strategies and engagement options available to the group as a whole and reduces hasty decisions. The engagement of multiple agents and multiple clients is different from a group of one-agent and one-client engagements. In existing concepts, the engagement space is simply partitioned by agents, but the possibility of using group engagement options is lost, and the problem of redefining the engagement space is likely to occur.

[0026] The approach of MAEDS varies depending on how each system handles the above-mentioned obstacles. The difficulty in making engagement decisions for multiple agents and avoiding hasty decisions is related to the agent information manageable in the above-mentioned MAEDS system. Making multiple engagement decisions adjusted in a timely manner and the difficulty of managing time and resources in automatically adjusting to new change drivers as the engagement evolves involve uncertainty and non-persistence, making it even more difficult. The MAEDS system has to judge the timing in order to adapt to such uncertainty and non-persistence and not lose sight of the engagement objectives.

[0027] Centralized System vs. Distributed SystemThere are two design approaches to the overall design of the MAEDS system. These two design approaches are a centralized system and a distributed system. The centralized MAEDS that coordinates the participation decisions for agents has fewer design difficulties compared to the decentralized coordination participation system. In "Flexible Multi-Agent Decision Making Under Time Pressure" by Sanguk Noh and Piotr J. Gmytrasiewicz (IEEE Transactions on Systems, Man and Cybernetics - Part A: Systems and Humans, Vol. 35, No. 5, September 2005, pp. 697-707), S. Noh and P. J. Gmytrasiewicz pointed out that when agents are coordinated to be distributed, each agent makes decisions based on data and assumptions about the data and actions of other agents. The distributed system requires increased computational resources to handle pre-planned models and profiles, as well as hierarchical trees of estimated probabilities, and becomes infeasible in time-critical situations.

[0028] Management of Uncertainty. When making involvement decisions for multiple agents and clients, uncertainty, non-persistence, and incomplete data pose specific challenges. In "Applied Decision Making With Fast-and-Frugal Heuristics" (Journal of Applied Research in Memory and Cognition 5, April 18, 2016, pages 215-231) by Sebastian Hafenbraedl, Daniel Waeger, Julian N. Marewski, and Gerd Gigerrenzer, S. Hafenbraedl, Daniel Waeger, Julian N. Marewski, and Gerd Gigerrenzer claim that "in uncertain situations, accurate judgments usually do not require great effort or complex strategies." They state that the classical rational (RAT) approach using Bayesian probabilities, complex modeling, optimization of cost estimation, and sophisticated weighting algorithms, as described by E. Tiantaphyllou, B. Shu, S. Nieto Sanchez, and T. Ray in "Multi-Criteria Decision Making: An Operations Research Approach" (Encyclopedia of Electrical and Electronics Engineering, Volume 15, 1998, pages 175-186 (Sum of weights, Multi-Attribute Decision Making issues)), is suitable for risk assessment based on complete information but not for applied decision-making using incomplete information. The fast-and-frugal heuristic (FAFH) approach, based on simple rules and thresholds, measurable environmental parameters, and decision-making capabilities, is more suitable in terms of decision-making intuition, speed, transparency, cost-effectiveness, and robustness.The components of this approach consist of search rules, stopping rules, and decision rules. In "Evidence accumulation in decision making: Unifying the “take the best” and the “rational” models," by Michael D. Lee and Tarrant D.R. Cummins (Psychonomic Bulletin & Review, 2004, 11(2), pp. 343-352), M. D. Lee and T.D.R. Cummins pointed out that one of the FAFHs known as the Take the Best (TTB) decision model "often yields results equal to or better than competing rational models in terms of decision-making accuracy when tested in real-world domains and makes decisions more quickly with fewer cognitive resources." These are the characteristics that real-time autonomous MAEDS needs to incorporate.

[0029] Another way to manage the uncertainty and non-persistence of the client environment is to ensure information updates and use short histories in decision-making. In "Stochastic Models of Evidence Accumulation In Changing Environments" by Alan Veliz-Cuba, Zachary P. Kilpatrick, and Kresimir Josic (SIAM Review, Vol. 58, May 2005), A. Veliz-Cuba, Z. P. Kilpatrick, and K. Josic pointed out that "in an uncertain and non-persistent environment, an ideal observer discounts prior evidence at a rate determined by the vagaries of the environment, and the dynamics of evidence accumulation are governed by the information obtained in an average environmental epoch." By frequently updating the past client history or relying on it for a short period, it is helpful for application in a dynamic environment.

[0030] Application of Strategies to the Environment。In "Evidence accumulation in decision making: Unifying the “take the best” and the “rational” models" by Michael D. Lee and Tarrant D.R. Cummins (Psychonomic Bulletin & Review, 2004, 11(2), pp. 343 - 352), M. D. Lee and T.D.R. Cummins view the TTB and RAT approaches as the two extremes of sequential sampling decision-making that vary according to various threshold levels of evidence required for decision-making. The threshold level used to end evidence accumulation can vary for the involved environment and for the adjustment of the MAEDS purpose.

[0031] Accumulation of Strong Evidence for Decision Making 。In "Sequential evidence accumulation in decision making: The individual desired level of confidence can explain the extent of information acquisition" by Daniel Hausmann and Damian Laege (Judgment and Decision Making, Volume 3, Issue 3, March 2008, pp. 229 - 243), D. Hausmann and D. Laege suggest that the stopping point for the accumulation of sufficient evidence for decision-making is related to the "desired level of confidence". In a dynamic involved environment, due to incomplete information, a hard stopping point can never be achieved. Fortunately, client involvement typically involves the agent / client proximity, so the agent's client recognition, its relative position and activities, and the client value judgment ability improve with proximity and time, and confidence increases in decision-making. Prolonging decision-making in a dynamic involved environment leads to better-quality information and better judgment.

[0032] Urgency and Timeliness In "Decisions in Changing Conditions: The Urgency-Gating Model" by Paul Cisek, Genevieve Aude Puskas, and Stephany El-Murr (Journal of Neuroscience, September 16, 2009, 29-(37), pages 11560-11571), P. Cisek, G. A. Puskas, and S. El-Murr claim that, in addition to the strength of the accumulated evidence, the urgency to make a choice also contributes to turning a judgment into an action. This point is where, in a changing engagement environment, the decrease in the agent's resources and the increase in the time pressure forcing an action intersect. The former is related to the agent's engagement ability, and the latter is related to the changes in the client's activities and the client's value. The timeliness of MAEDS responds to emergency events.

[0033] Described below is a real-time MAEDS that provides timely adjusted simultaneous engagement decisions, and the timely adjusted simultaneous engagement decisions lead to the agent's energy conservation, engagement with the same number of high-value clients as the agent, reliable support for the highest-value clients, flexible engagement changes adapting to the development of client value, engagement opportunities, resource limitations, the influence of past service judgments, the activities and states of the agent and the client, and the urgency of engagement.

[0034] Figure 1 shows an aspect of MAEDS 110 in a centralized adjusted agent system 100, from agent 1021 to 102 N(Collectively, agents 102) communicate in real time with a central computer 106 via a communication system 104. Each agent 102 has a sensor system, a navigation system, a guidance control unit, a motion engine, and a communication system. The positions, activities, and statuses of agents 102 and clients are communicated to the agent and client status processor 108 of the central computer 106 via the communication system 104. The central computer 106 determines the client value at each point in time and provides a ranked list of clients to the MAEDS 110. The central computer 106 may be hosted externally or may be hosted by one of the agents 102.

[0035] FIG. 6 is a block diagram showing elements of a multi-agent engagement decision system (MAEDS) 110 configured to implement engagement decisions according to some aspects of the concepts of the present invention. As shown, the MAEDS 110 may include a network interface 607 and a transceiver circuit 601 including a transmitter and a receiver configured to provide uplink and downlink wireless communication with agents of the network. The MAEDS 110 may further include a processing circuit 603 (also referred to as a processor) coupled to the transceiver circuit and a memory circuit 605 (also referred to as a memory) coupled to the processing circuit. The memory circuit 605 may include computer-readable program code. When executed by the processing circuit 603, this computer-readable program code causes the processing circuit to perform the operations according to the embodiments described herein. According to other embodiments, the processing circuit 603 may be defined to include the memory, thereby eliminating the need for a separate memory circuit. The MAEDS 110 may further include an interface (e.g., a user interface) coupled to the processing circuit 603 and / or the MAEDS 110 may be incorporated within a vehicle.

[0036] As described above, the operation of MAEDS110 may be executed by the processing circuit 603 and / or the transceiver circuit 601. For example, the processing circuit 603 controls the transceiver circuit 601 to transmit communication content to the agent 102 through the transceiver circuit 601 via a wireless interface, and / or receive communication content from the agent 102 through the transceiver circuit 601 via the communication system 104. Further, the modules may be stored in the memory circuit 605. These modules can provide instructions, and when the instructions of the modules are executed by the processing circuit 603, the processing circuit 603 executes the corresponding operations.

[0037] Returning to FIG. 1, the agent 102 can determine its own position, move, and communicate information to the central computer 106. The agent 102 can monitor the client environment. A "client" refers to an important customer. This client requires all the attention of the agent 102, and in some embodiments, the agent may not provide services to any other client later. For the client group, each member is designated as an important client. The tasks of the agent include obtaining data regarding the value and needs of the client and providing information to the MAEDS110 via the agent and the client status processor 108.

[0038] MAEDS 110 outputs the participation decision parameters used by the look processor 112 and the motion processor 114. The look processor 112 receives the participation decision parameters via the communication system 104 and provides a look vector for each agent 102. The motion processor 114 receives the participation decision parameters via the communication system 104 and provides a motion vector for each agent 102. The agent 102 receives the look vector and the motion vector, updates the guidance control unit based on the look vector and the motion vector, and controls the motion engine accordingly.

[0039] MAEDS 110 operates with multiple agents 102. Below, the participation options of MAEDS are described as the possible sets of agent / client pairs for participation. MAEDS participation decision refers to the participation action plan after selecting the best participation option and reviewing the necessary actions.

[0040] An advantage that can be realized by using MAEDS 110 is the real-time and novel manner of MAEDS 110 that makes decisions autonomously based on the activities and status of the latest agents and clients. MAEDS 110 does not require prior planning and does not estimate the number of clients or healthy agents in the environment. The real-time approach enables reconsideration of participation decisions, resulting in flexible participation decisions and adaptation to deployment in service scenarios.

[0041] Another achievable advantage is a novel manner of the participation decision process using a prioritized client "tier" approach. By using a tier approach starting from the first tier, MAEDS 110 ensures that, if possible, participation decisions are made for the highest-ranked clients.

[0042] Another advantage of using MAEDS110 is that the criteria for new engagement selection are based on "urgency" and "risk". Therefore, the permission, timeliness, and risk of making an engagement decision can be based on the engagement strategy of MAEDS and the agent's ability to complete the engagement for the client within the hierarchy being analyzed.

[0043] Another achievable advantage is a new engagement option evaluation that depends on converting the client's value rank into a value matrix of powers of two integers. Unlike traditional weighting schemes that proportionally tie weights to client values, MAEDS110 focuses on relative values (ranks) rather than absolute values. In an environment where client values fluctuate, client ranks are more stable, providing more stable results by minimizing dependence on absolute client values.

[0044] Another achievable advantage is that MAEDS simultaneously solves the problems of multiple agent / client pairs, thereby avoiding interference with engagement options due to previous sequential customer engagement decisions or agent and client cluster area commitments. In the evaluation of engagement options for all possible agent / client pairs, high-rank client values, adjusted based on the engagement ability, urgency, and risk of the agent / client pair to successfully complete the service, are prioritized.

[0045] Referring to FIG. 2, a high-level overview of the MAEDS function is shown. MAEDS 110 receives the agent motion and status of each agent, the ranked client list of the client, and the motion of each client. At block 201, MAEDS 110 determines the hierarchy. At block 203, MAEDS 110 determines the permissions, urgency, and risks. At block 205, MAEDS determines the participation options. At block 207, MAEDS 110 makes a participation decision and sends the participation parameters to the agent. MAEDS repeats blocks 201, 203, 205, and 207 for each hierarchy. The central computer 106 periodically updates the agent motion and status of each agent, the ranked client list of the client, and the motion of each client. MAEDS 110 repeats blocks 201, 203, 205, and 207 for each update.

[0046] Referring to FIG. 3, further details of the process flow of MAEDS in FIG. 2 are shown. At block 301, the MAEDS process starts and the number of hierarchies is determined. Based on the number of agents that require a participation decision, multiple hierarchies are determined by dividing the ranked client list into hierarchies. At block 303, MAEDS determines the number of agents and clients in the hierarchy under analysis.

[0047] Based only on the rank of the client values, clients are grouped into "tiers", and the number of clients in each tier is equal to the number of agents that need to make decisions. The first tier consists of the highest-ranked clients, followed by the second tier of the next highest-ranked clients, and so on. The last tier includes the remaining clients, and the number may or may not be equal to the number of agents that need to make decisions. By starting from the first tier and checking the client list, MAEDS110 ensures that, if possible, an engagement decision is made for the highest-ranked clients. Otherwise, an engagement decision is made for the clients in the next tier. This ensures that MAEDS110 mobilizes all agents to engage with clients as much as possible.

[0048] This tiered feature allows for options for over-engagement when more than one agent is required to engage with a client to provide services. This strategy is typically used when there are more agents than clients. In this feature, using tiers allows for some decisions for over-situations without making the MAEDS decision of which agent to select more complex. Since there are more agents than clients in the tier, there is only one tier. MAEDS110 places the over-abundant clients in the tier at the end of the client list in the tier, virtually increasing the number of clients that an agent can match. This strategy is easily adapted in the rest of the MAEDS processing, and the best over-engagement options and decisions are determined in a timely manner.

[0049] In block 305, MAEDS110 calculates an engagement ability parameter for each agent / client pair. The permission, timeliness, and risk to make an engagement decision are based on the MAEDS engagement strategy and the ability of the agent to complete the engagement for the clients in the tier.

[0050] The margin of agent involvement possibility and the start time to complete the service margin are typical involvement capacity parameters used for the determination of the permission, timeliness, and risk of each agent / client pairing within the hierarchy. The proximity to the client margin is also sometimes used. The selection of the involvement capacity parameters depends on the purpose of the involvement.

[0051] If the involvement capacity margin of an agent / client pair is positive, that agent / client pair is "permitted" in the consideration of involvement options. Otherwise, it is "rejected".

[0052] In block 307, MAEDS110 determines whether there are urgent agent / client pairs within the hierarchy and whether there are risky agent / client pairs within the hierarchy. In block 309, MAEDS110 stores the identifiers (IDs) of the urgent agent / client pairs and the risky agent / client pairs. When the involvement capacity of an agent / client pair is "urgent", the urgent criteria are set and established. When the involvement capacity margin is within the framework of the upper and lower threshold criteria, the agent / client pair is declared "urgent". By checking the current and next involvement capacity margins, the timeliness of the urgent declaration is ensured. As described above, the urgency of the agent / client pair is used to adjust the value of the involvement option.

[0053] In some aspects, the design of the urgent criteria depends on the strategy of the MAEDS used. The strategy of the MAEDS can vary depending on whether the agent ends after one involvement or continuously moves on to other involvements. The urgent criteria can be used to adjust for strategic differences.

[0054] Even if a particular agent / client pair is permitted and is urgent, the engagement capacity margin for that agent / client pair may still be riskier than other agent / client pairs within the same hierarchy. This is considered in light of possible future dynamic events (e.g., fluctuating client value, unexpected client activities, unexpected agent health status), and with a low engagement capacity margin, it may not be possible to support service completion. When the engagement capacity margin of an agent / client pair falls below an acceptable threshold, that agent / client pair is assigned a non-zero risk level. In one aspect, the risk level is represented by a fractional risk value between 0 and 1 that is proportional to the lack of the desired engagement capacity margin. The design of the risk value depends on the MAEDS engagement strategy, agent type, and engagement capacity used. Thus, other risk level values may be used.

[0055] Table 1 outlines permission, urgency, and risk level determination. Table 1 Outline of Permission, Urgency, and Risk TIFF0007713813000001.tif69170

[0056] MAEDS engagement options are a set of possible agent / client pairings for a particular engagement. MAEDS engagement options are determined within the hierarchy of agents that require an engagement decision. If there are agents remaining in a hierarchy that require an engagement decision, these agents are carried over, if available, to be used in the next client hierarchy.

[0057] In some aspects of the concepts of the present invention, the client rank is converted to a power of two integer value within the value matrix. The value matrix is used in the determination of the participation option. Unlike conventional weighting schemes that proportionally associate weight values with client values, MAEDS focuses on relative values (ranks) rather than absolute values. In an environment where client values vary, the client rank is more stable and provides more stable results by minimizing dependence on absolute client values.

[0058] In converting the client rank to a power of two integer value, the value of the client with the lowest rank within the hierarchy is 1. As the client's rank increases, its value increases by the next power of two. Thus, the client ranks for four client rankings are 1, 2, 4, and 8.

[0059] In block 311, MAEDS 110 sets the hierarchical value matrix. In block 313, MAEDS 110 modifies the value matrix for non-participating agent / client pairs. Exemplary value matrices for four agents and four clients within the hierarchy are shown in FIGS. 4a and 4b. FIG. 4a shows an exemplary value matrix where all agent / client pairs are permitted. FIG. 4b shows an exemplary value matrix where four agent-client pairs (A1C3, A2C4, A3C1, and A4C4) are rejected.

[0060] In block 305, a participation ability parameter is calculated for each agent / client pair. Using the participation ability margin of the agent / client pair, the permission of the agent / client pair is determined. If the ability margin exceeds the minimum threshold, it is permitted; otherwise, it is rejected. The permitted pairs retain their values within the value matrix. The rejected pairs are given a value of 0 within the matrix. In one aspect, only clients having at least one permitted agent / client pair within the value matrix are used.

[0061] During the process of participation, a situation may occur where there are no permissible clients for an agent to pair with within a hierarchy. An example of the value matrix for such a case is shown in Figure 4c. Here, Agent 3 has no client to pair with. In the example of Figure 4c, Agent 3 is retained to be used with clients in the next hierarchy (if any exist). If there are no clients left, Agent 3 uses the default decision. The default decision is to continue monitoring for an opportunity to participate without making any changes to its course.

[0062] More examples of agents and clients are shown in Figure 4d. In this example, there are four agents that require participation decisions, but only two clients. In Figure 4d, there is only one hierarchy for establishing participation options.

[0063] The value matrix in Figure 4d may be evaluated as is, or it may be adjusted to achieve agent overload for the clients before evaluation.

[0064] The setting for overloading agents for the clients is shown in Figure 4e. Here, the third and fourth rows of the value matrix correspond to the data of the two clients and have values related to the third and fourth ranked clients within the hierarchy. Since there are effectively four clients, the original two clients will obtain higher powers of two values.

[0065] Overload means the possibility of a commitment to invest more resources to engage with the same client. Therefore, caution must be exercised when overloading. In an environment where client value changes dynamically, the risk of overloading agents for clients too early can be reduced by delaying the overload until the number of clients in that scenario is certain, or by not fully overloading agents for clients (for example, repeating only client 1 in the third row instead of repeating both clients).

[0066] MAEDS110 simultaneously solves the problems of multiple agent / client pairs, thereby avoiding interference with engagement options due to previous sequential customer engagement decisions or agent and client cluster area commitments. Using all possible combinations of agent / client pairs, engagement options are evaluated, thereby making an adjusted solution and solving the difficulty of differentiating overlapping areas of client engagement capabilities.

[0067] The evaluation of all possible engagement options is performed using the Greedy Knapsack algorithm (~O: N log N in computational load). Using integer values in the value matrix minimizes the computational load. Each engagement option is a path from the top to the bottom of the value matrix and consists of groups of individual agent / client pairs. An agent cannot pair with more than one client, and a client cannot pair with more than one agent (except in the case of overload). The path value (PV) is the sum of the values of the permissible agent / client pairs involved in the engagement option.

[0068] The advantage of the mode of converting the above client rank into a power of 2 integer value is that the PV uniquely reveals the clients involved in the participation option (due to the unique binary characteristics of the power of 2). As a result, the PV is neatly restricted to an integer value with known upper and lower limits.

[0069] Figure 4f shows three exemplary pair paths of the value matrix using the matrix of Figure 4b. That is, three participation options are shown. Path 401 (striped arrow) shows agent / client pairs A1C1, A3C2, A4C3, and A2C4. Path 403 (dotted arrow) shows agent / client pairs A3C1, A4C2, A1C3, and A2C4. Path 405 (solid arrow) shows agent / client pairs A1C1, A2C2, A4C3, and A3C4. The initial PV of path 401 is 8 + 4 + 2 + 0 = 14. The initial PV of path 403 is 0 + 4 + 0 + 0 = 4. The initial PV of path 405 is 8 + 4 + 2 + 1 = 15.

[0070] In the case of multiple tiers where the agent is carried over to the next tier, it is possible that there are only dropped value inputs in the client rows. To reflect this condition, a skip client display is used. Under the skip client condition, the agent has the freedom to skip a client from consideration for the participation option. For the agent paired with that client, a matrix value of 0 is used.

[0071] Returning to Figure 3, in block 315, MAEDS110 finds one or more pair paths that initially have the highest total value among the pair paths. In the path with the highest total value, there may be more than one pair path. For example, in Figure 4f, the pair paths that initially have the highest total value are the pair paths with a path value of 15. These pair paths are the pair paths having the following agent / client pairs. A1C1, A2C2, A4C3, A3C4 A1C1, A3C3, A2C3, A3C4 A2C1, A1C2, A4C3, A3C4 A2C1, A3C2, A4C3, A1C4 A2C1, A4C2, A3C3, A1C4 A4C1, A1C2, A2C3, A3C4 A4C1, A2C2, A3C3, A1C4 A4C1, A3C2, A2C3, A1C4

[0072] The general operation that MAEDS performs to find all paths with the highest initial point value is as follows. 1. Set up a value matrix for the hierarchy using powers of 1.2 for value display. 2. For any agent / client pair that is rejected, set the matrix element to 0. 3. If desired or possible, adjust to overload agents for clients. 4. Agents that create a list of agents with all rejected clients within the hierarchy are retained for use with clients in the next layer, if any. 5. Execute a greedy knapsack on the value matrix a. For each agent, find a matching client. Consider the possibility that the client has no permissible agents with which to match. When such a situation occurs, mark it as a skip client. b. Number each possible path. c. Calculate the path value (PV) by summing the values of the matrix elements within the path. d. Rank the paths by PV. e. Find all paths with the same highest PV.

[0073] The pair path with the highest initial PV is displayed as a candidate pair path. If multiple candidate paths have the same highest PV (such as the example in Figure 4f), the urgency of the agent / client pair's engagement ability is used to narrow down the best candidate path for the engagement decision.

[0074] The urgency condition of the agent / client pair can be used to prevent the engagement decision from being made prematurely by lowering the PV value of the candidate pair path when there is an urgent agent / client pair within the candidate pair path. If the candidate pair path has a non-urgent agent / client pair, the default decision is for the agent of the agent / client pair to stay on course without taking any action. Even if the pair is urgent, if there are other candidate pair paths that do not include that pair, that pair is not urgent with respect to the engagement decision of MAEDS. In down-selection, paths without urgent pairs are prioritized. This avoids making the engagement decision prematurely based on the urgency of individual pairs.

[0075] For any candidate path with an engagement ability lower than the desired minimum value, a further decrease in the PV value is made. This reflects the risk of the candidate path in completing the engagement in view of the client's needs or possible future fluctuations with unexpected changes. The risk value in one example is a fraction between 0 and 1.

[0076] In block 317, MAEDS 110 lowers the value of the pair path having an urgent agent / client pair and / or a risky agent / client pair. In block 319, the best path is selected. By this process of lowering the value, it is ensured that the candidate path with the highest PV after lowering the value, which indicates the best path, has the agent most likely to have appropriate resources for serving the client. In case of a tie, the first initially highest PV candidate path is the best path.

[0077] In block 321, MAEDS110 determines whether the best path has an urgent agent / client pair. If the best path includes an urgent agent / client pair, the agent must make an engagement decision to take action; otherwise, it is decided to stay on course by default and no action is taken. This feature saves agent resources for broader scenario development during the service process. Therefore, when the decision condition is met in block 323, an engagement decision is made in block 325.

[0078] If the decision condition is not met in block 323, MAEDS110 determines in block 327 whether only one agent can engage with an urgent client. If only one agent can engage with an urgent client in block 327, an engagement decision is made in block 325.

[0079] If there is no urgent agent / client pair in the best path, or if MAEDS110 makes an engagement decision, MAEDS110 proceeds to block 329. If the decision condition is not met in block 323, MAEDS110 also proceeds to block 329, where more than one agent can engage with an urgent client.

[0080] In block 329, MAEDS110 determines whether there are any agents in the hierarchy that cannot engage with the client. If there are no agents in the hierarchy that cannot engage with the client, MAEDS110 generates engagement parameters in block 333. If, as determined in block 331, there are no more hierarchies to analyze, MAEDS110 also generates engagement parameters. If there are agents in the hierarchy that cannot engage with the client and there are more hierarchies as determined in block 331, MAEDS proceeds to analyze the next hierarchy and repeats blocks 303 to 333 for each layer.

[0081] If there are more agents that require a decision, these agents are used for the next level of clients. This ensures that each agent has a chance to engage with the most likely client needs in the client needs list. This is because for a given agent / client pair, a change in the rank of the fluctuating client value will lower the rank of a client that was previously ranked highly, and MAEDS ensures that all agents are used even if it had previously made a decision for an agent to engage with that client.

[0082] To find the decision parameters to send to an agent, the general operation that MAEDS performs is as follows.

[0083] 1. For all paths having the same highest value, indicate these as candidate paths. a. Modify the candidate path value i. Any candidate path that also has an urgent agent / client pair decreases in PV by a power of 2 of the client's value in the value matrix. ii. The PV of the candidate path decreases by only the risk value. c. Find the best path. The best path is the first path having the highest adjusted PV. 2. Make a decision a. If the best path contains an urgent pair, the number of agents is more than 1, or the number of agents is equal to 1 and the engagement ability is within the decision threshold, or the start time to complete the service is within the time threshold, decide for the agent within the urgent pair to act. b. Determine the appropriate action for the agent within the urgent pair. The action depends on the type of agent involved. This typically involves whether the agent moves or performs an engagement action. c. The decision for the agent within the best non-urgent path is to stay on course without taking any action. 3. If there are agents for which a decision is required and which cannot interact with any of the clients within a hierarchy, the process described above is repeated at the next hierarchy. 4. Generate decision parameters for output to the agent.

[0084] Having thus described the overall operation of MAEDS, an example of the operations performed by MAEDS 110 will now be described. From here on, the operations of MAEDS 110 (the operations implemented using the block diagram structure of FIG. 6) will be described with reference to the flow diagrams of FIGS. 7a and 7b, which relate to some aspects of the concepts of the present invention. For example, the modules may be stored in the memory 605 of FIG. 6. These modules can provide instructions, and when the instructions of the module are executed by the corresponding MAEDS processing circuit 603, the processing circuit 603 executes the corresponding operations of the flow diagram.

[0085] Looking at FIG. 7a, at block 701, the processing circuit 603 determines agent / client pairs in at least one priority hierarchy based on the number of clients and the number of agents, and each agent / client pair has an agent, a client, and a client's client rank.

[0086] FIG. 8 shows an aspect of determining an agent / client pair. At block 801, the processing circuit 603 dynamically obtains from the status processor the number of clients, the number of agents, and for each client, the client's client rank. For example, in relation to FIG. 1, the processing circuit 603 dynamically obtains from the agent and client status processor 108 the number of clients, the number of agents, and for each client, the client's client rank.

[0087] In block 803, the processing circuit 603 determines at least one priority level based on the number of clients, the client rank of each client, and the number of agents. In block 805, the processing circuit 603 groups some agent / client pairs into at least one priority level according to the client rank of the clients in the agent / client pairs.

[0088] Returning to FIG. 7a, in block 703, the processing circuit 603 determines the matrix value of each agent / client pair according to the client rank of the clients in the agent / client pairs. FIG. 5a shows an example of the matrix value of each agent / client pair based on the client rank of the clients in the agent / client pairs. This value matrix is an example from FIG. 4b.

[0089] Referring to FIG. 9, the processing circuit 603 of block 901 can derive a value matrix of at least one priority level based on the number of agents and the number of clients in at least one priority level. This value matrix has a number of rows that is less than or equal to the number of agents for which a decision is required and less than or equal to the number of clients. In block 903, the processing circuit 603 enables more than one agent to be involved with a client by adding rows with lower rankings to the value matrix for the client.

[0090] Referring to FIG. 10, when a power of 2 is used in the value matrix, the processing circuit 603 of block 1001 sets the matrix value of each agent / client pair in the bottom row of the value matrix to a value of 1. The processing circuit 603 of block 1003 sequentially sets the matrix value of each agent / client pair having a higher client rank to a value that is sequentially higher by a power of 2 N-1 . N is the number of rows in which the agent / client pairs are arranged within the value matrix, and the matrix value of the bottom row is 1.

[0091] Returning to FIG. 7a, at block 705, the processing circuit 603 determines the engagement capability parameters for each agent / client pair. For example, the start time for service completion for all agent / client pairs is given sufficient leeway so as not to be a factor affecting urgency. Let E be the engagement capability of an agent / client pair. In this example, E is related to the urgency of the agent. EM represents the amount of energy margin for service completion. An example of the engagement capability margin is shown in FIG. 5b.

[0092] At block 707, the processing circuit 603 sets the matrix value for each rejected agent / client to a zero value. A rejected agent / client pair has an engagement capability margin less than a predetermined threshold. For example, less than the predetermined threshold can be zero. In other words, the engagement capability margin must be a positive value. Thus, the matrix value of each agent / client pair having a negative value is set to zero. In FIG. 5b, the engagement capability margins of the agent / client pairs A1C3, A2C4, A3C1, and A4C4 are negative values. Thus, as shown in FIG. 5a, the matrix values of the agent / client pairs A1C3, A2C4, A3C1, and A4C4 are set to zero.

[0093] In block 709, the processing circuit 603 evaluates a plurality of participation options. Each participation option is a pair path from the agent / client pair with the highest client rank to the agent / client pair with the lowest client rank, and each pair path has an agent / client pair for each client in at least one priority level. FIG. 4f shows three pair paths. In an embodiment, the value matrix of FIG. 5a is a pair path from the agent / client pair in the top row of the value matrix to the agent / client pairs in each middle row of the value matrix and to the agent / client pair in the bottom row of the value matrix, and each agent and client of each agent / client in the pair path is different from the other agents and clients of the agent / client pair of the pair path.

[0094] In block 711, the processing circuit 603 calculates an initial path value for the pair path of each participation option. An example of calculating the initial path value for three pair paths was described above in the description of FIG. 4f.

[0095] In block 713, the processing circuit 603 determines a candidate pair path by determining the pair path having the highest initial path value. The pair path having the highest initial path value is the candidate pair path. In the exemplary value matrix of FIG. 5a, the highest initial path value is 15. There may be a plurality of pair paths having the highest initial path value.

[0096] In block 715, the processing circuit 603 reduces the initial path value of each candidate pair path based on the agent / client pair in the candidate pair path that is the critical agent / client pair and based on the agent / client pair in the candidate pair path that is the risky agent / client pair in order to derive the final path value for the candidate pair path.

[0097] Referring to FIG. 11, the processing circuit 603 of block 1001, in block 1101, in response to the engagement capability parameter of the agent / client pair being within the range of the first threshold and the second threshold, determines that the agent / client pair is an emergency agent / client pair, and in block 1103, in response to the engagement capability parameter of the agent / client pair being less than the risk threshold, determines that the agent / client pair is a risky agent / client pair, thereby determining an agent / client pair that is an emergency agent / client pair and an agent / client pair that is a risky agent / client pair.

[0098] For example, in relation to FIG. 5a, the first threshold may be zero and the second threshold may be 65. A risky agent / client pair is set to be any agent / client pair whose engagement capability margin is less than the risk threshold of 35. Thus, any agent / client pair in FIG. 5a whose engagement capability margin is less than 65 (i.e., between 0 and 65) is an emergency agent / client pair, and any agent / client pair whose engagement capability margin is less than 36 is a risky agent / client pair. In the example, since the risk threshold is between the first threshold and the second threshold, a risky agent / client pair is also an emergency agent / client pair. In FIG. 5a, the emergency agent / client pairs are the A4C2, A3C3, A4C3, and A1C4 agent / client pairs. The risky agent / client pairs are the A1C4 and A4C3 agent / client pairs.

[0099] Referring to FIG. 12, for each agent / client pair in the candidate pair path that is an emergency agent / client pair, the processing circuit of block 1201 decreases the initial path value of the high-value pair path by the matrix value of the emergency agent / client. In block 1203, the processing circuit 603 decreases the initial path value of the candidate pair path for each agent / client pair in the candidate path that is a risky agent / client pair, based on the participation ability parameter of the risky agent / client pair.

[0100] In the example of FIG. 5a, the risk value, which is a fractional value, is defined by Table 2.

[0101] Table 2 Examples of Risk Values TIFF0007713813000002.tif38170

[0102] In this example, assume that for all agent / client pairs from one MAEDS cycle to the next, the expected decrease in EM is 1. Therefore, it is not expected that the permission, urgency, and risk of any participation option will change within the MAEDS cycle.

[0103] The participation option paths with the corresponding agent / client pairs, initial PV, rounded-down urgency values, and risk values are shown in Table 3. The cells with dotted shading are emergency agent / client pairs, and the cells with vertical stripes are emergency and risky agent / client pairs.

[0104] Table 3 Participation Options and Values TIFF0007713813000003.tif177170

[0105] Sorted by the initial PV, the paths with the highest PV of 15 are shown. These are the candidate paths shown in Table 4. Further, Table 4 shows the final PV after reducing the value by urgency and risk.

[0106] Table 4 Candidate pair paths and PV TIFF0007713813000004.tif58170

[0107] In block 717, the processing circuit 603 determines the best path based on the final path value for each candidate pair path. The best path has a path number of 19 and the highest final PV 15. This path, in addition to being the best path, will further include all agent / client pairs having the highest EM. In some embodiments, if there is more than one candidate pair path having the highest final PV, the first candidate pair path in the list having the highest final PV is selected as the best path. Other ways to select the best path when there is more than one candidate pair path having the highest final PV include selecting a candidate path having no urgent agent / client pairs and / or risky agent / client pairs, selecting a candidate path having the fewest number of urgent agent / client pairs, selecting a candidate path having the fewest number of risky agent / client pairs, selecting a random candidate path, and the like.

[0108] In block 719, the processing circuit 603 derives at least one involvement decision based on the best path. At path number 19, since there are no urgent agent / client pairs, there is no need to commit the agent to involve any client. The involvement decision to be made is that all agents remain on their course without change until the next MAEDS cycle.

[0109] In block 721, the processing circuit 603 transmits at least one involvement decision to the agents in the best path.

[0110] In the above example, there was no emergency agent / client pair on the best path. This is not always the case. Referring to FIG. 13, when there is an emergency agent / client pair on the best path, the processing circuit 603, in response to the best path having an emergency agent / client pair, the engagement ability of the emergency agent / client pair being within a decision threshold, or the start time being within a time threshold, derives at least one engagement decision based on the best path. At block 1301, the first action to be performed by the agent of the emergency agent / client pair is determined. At block 1303, in response to determining the first action, the processing circuit 603 transmits the first action to the agent of the emergency agent / client pair.

[0111] At block 1305, the processing circuit 603 determines the second action to be taken by the agent of the agent / client pair on the best path that is not the emergency agent / client pair, in response to the agent of the agent / client pair on the best path. The second action to be taken by the agent of the agent / client pair on the best path is to stay on course. At block 1307, the processing circuit 603 transmits the second action of staying on course to the agent of the agent / client pair on the best path that is not the emergency agent / client pair.

[0112] The scope of protection is determined by the scope of the appended claims. However, the understanding of the present disclosure can be obtained in various ways including, but not limited to, the following clauses.

[0113] Clause 1 A method of generating a set of adjusted engagement parameters by a processor in a multi-agent engagement decision system, comprising: Determining agent / client pairs in at least one priority level based on the number of clients and the number of agents, wherein each agent / client pair has an agent, a client, and the client rank of the client, and determining the agent / client pairs; Determining a matrix value for each agent / client pair based on the client rank of the client of the agent / client pair; Determining a participation ability parameter for each agent / client pair; Setting the matrix value for each rejected agent / ranked client value to a zero value, wherein the rejected agent / client pair has a participation ability margin less than a predetermined threshold, and setting it to the zero value; Evaluating a plurality of possible participation options, wherein each participation option is a pair path from the agent / client pair with the highest client rank to the agent / client pair with the lowest client rank, and each pair path has an agent / client pair for each client in the at least one priority level, and evaluating the plurality of possible participation options; Calculating an initial path value for each pair path of the participation option for each participation option; Determining a candidate pair path by determining the pair path having the highest initial path value, wherein the pair path having the highest initial path value is the candidate pair path, and determining the candidate pair path; For each candidate pair path, reducing the initial path value of the candidate pair path based on the agent / client pair in the candidate pair path that is an urgent agent / client pair and based on the agent / client pair in the candidate pair path that is a risky agent / client pair, in order to derive a final path value for the candidate pair path; Determining an optimal path based on the final path values of the candidate pairs of paths; Deriving at least one participation decision based on the optimal path; Sending the at least one participation decision to an agent within the optimal path; A method comprising.

[0114] Clause 2 In response to the participation ability parameter of the agent / client pair being within a range of a first threshold value and a second threshold value, determining that the agent / client pair is an emergency agent / client pair; In response to the participation ability parameter of the agent / client pair being less than a risk threshold value, determining that the agent / client pair is a risky agent / client pair; The method according to clause 1, further comprising determining an agent / client pair that is an emergency agent / client pair and an agent / client pair that is a risky agent / client pair.

[0115] Clause 3 Determining an agent / client pair in the at least one priority level based on the number of the clients and the number of the agents; Dynamically obtaining, from a status processor, the number of the clients, the number of the agents, and for each client, the client rank of the client; Determining the at least one priority level based on the number of the clients and the number of the agents; Grouping some agent / client pairs into the at least one priority level according to the client rank of the clients in the agent / client pairs; The method according to clause 1 or 2, comprising.

[0116] Clause 4 Deriving a value matrix for the at least one priority level based on the number of agents and the number of clients in the at least one priority level, the value matrix having a number of rows less than or equal to the number of agents for which a decision is required and less than or equal to the number of clients, the method according to any one of clauses 1 to 3 further comprising deriving the value matrix for the at least one priority level.

[0117] Clause 5 The method according to clause 4, wherein deriving the value matrix further comprises adding lower-ranked rows to the value matrix for the clients to enable more than one agent to be involved with a client.

[0118] Clause 6 Each participation option is a pair path from the agent / client pair in the top row of the value matrix to the agent / client pairs in each intermediate row of the value matrix and to the agent / client pair in the bottom row of the value matrix, each agent and client in each agent / client in the pair path being different from the other agents and clients in the agent / client pair of the pair path, the method according to clause 4 or 5.

[0119] Clause 7 Setting the matrix value of each agent / client pair in the bottom row of the value matrix to a value of 1, and Setting the matrix value of each agent / client pair having successively higher client ranks to successively higher values by a power of 2 N-1 and, N being the number of rows in which the agent / client pair is located within the value matrix, the bottom row matrix value being 1, the method according to any one of clauses 4 to 6.

[0120] ​ Clause 8 Based on the agent / client pairs in the high-value pair path that are emergency agent / client pairs and based on the agent / client pairs in the high-value pair path that are risky agent / client pairs, reducing the initial path value of the candidate pair path is, For each agent / client pair in the candidate pair path that is an emergency agent / client pair, reducing the initial path value of the candidate pair path by only the matrix value of the emergency agent / client, For each agent / client pair in the candidate path that is a risky agent / client pair, reducing the initial path value of the candidate pair path by a risk value based on the participation ability parameter of the risky agent / client pair The method according to clause 7, comprising.

[0121] Clause 9 Deriving the at least one participation decision based on the best path is, In response to the best path having an emergency agent / pair and the participation ability of the emergency agent / pair being within a decision threshold or the start time being within a time threshold, determining a first action to be performed by the agent of the emergency agent / client pair, Determining a second action to be performed by the agent of the agent / client pair in the best path in response to the agent of the agent / client pair in the best path that is not an emergency agent / client pair, the second action to be performed by the agent of the agent / client pair in the best path being to stay on course, Including, Sending the at least one participation decision to the agent in the best path is, In response to determining the first action, transmitting the first action towards the agent of the emergency agent / client pair Transmitting the second action of staying on the course towards the agent of the agent / client pair in the best path that is not an emergency agent / client pair The method according to any one of clauses 1 to 8, comprising:

[0122] Clause 10 A computer program having a non-transitory computer-readable medium with computer-executable instructions, which, when executed by a processor included in a device, cause the device to Determine agent / client pairs in at least one priority level based on the number of clients and the number of agents, each agent / client pair having an agent, a client, and a client rank of the client; Determine a matrix value for each agent / client pair based on the client rank of the client of the agent / client pair; Determine a participation ability parameter for each agent / client pair; Set the matrix value to a zero value for each discarded agent / ranked client value, where the discarded agent / client pair has a participation ability margin less than a predetermined threshold; Evaluating a plurality of possible engagement options, each engagement option being a pair path from the agent / client pair with the highest client rank to the agent / client pair with the lowest client rank, and each pair path having an agent / client pair for each client in the at least one priority level, and evaluating a plurality of possible engagement options; For each engagement option, calculating an initial path value for each pair path of the engagement option; Determining a candidate pair path by determining the pair path having the highest initial path value, wherein the pair path having the highest initial path value is the candidate pair path; For each candidate pair path, reducing the initial path value of the candidate pair path based on the agent / client pair in the candidate pair path that is the critical agent / client pair and based on the agent / client pair in the candidate pair path that is the risky agent / client pair, in order to derive a final path value for the candidate pair path; Determining the best path based on the final path values of each candidate pair path; Deriving at least one engagement decision based on the best path; Transmitting the at least one engagement decision to the agent within the best path; A computer program for causing a computer to execute operations including the above.

[0123] Clause 11 The non-transitory computer-readable medium further comprises additional computer-executable instructions that, when executed, cause the device to Determine that an agent / client pair is a critical agent / client pair in response to the engagement ability parameter of the agent / client pair being within a range of a first threshold and a second threshold; In response to the engagement ability parameter of the agent / client pair being less than the risk threshold, the agent / client pair determines that it is a risky agent / client pair determining an agent / client pair that is an emergency agent / client pair and an agent / client pair that is a risky agent / client pair, The computer program according to clause 10, which causes an operation including this to be executed.

[0124] Clause 12 determining an agent / client pair in the at least one priority level based on the number of the clients and the number of the agents, dynamically obtaining, from a status processor, the number of the clients, the number of the agents, and for each client, the client rank of the client, determining the at least one priority level based on the number of the clients and the number of the agents, grouping several agent / client pairs into the at least one priority level according to the client rank of the client in the agent / client pair The computer program according to clause 10 or 11, including this.

[0125] Clause 13 The non-transitory computer-readable medium comprises further computer-executable instructions that, when executed, cause the device to derive a value matrix for the at least one priority level based on the number of agents and the number of clients in the at least one priority level, the value matrix having a number of rows less than or equal to the number of agents for which a decision is required and less than or equal to the number of clients, and cause the device to execute an operation including deriving the value matrix for the at least one priority level, the computer program according to any one of clauses 10 to 12.

[0126] Clause 14 The non-transitory computer-readable medium comprises further computer-executable instructions that, when executed, cause the device to execute an operation including enabling more than one agent to participate in the client by adding a lower-ranked row to the value matrix for the client, the computer program according to clause 13.

[0127] Clause 15 Each participation option is a pair path from the agent / client pair in the top row of the value matrix to the agent / client pairs in each middle row of the value matrix and to the agent / client pair in the bottom row of the value matrix, each agent and client of each agent / client in the pair path being different from the other agents and clients of the agent / client pair of the pair path, the computer program according to clause 13 or 14.

[0128] Clause 16 The non-transitory computer-readable medium comprises further computer-executable instructions that, when executed, cause the device to Setting the matrix value of each agent / client pair in the bottom row to a value of 1, and for each agent / client pair having a successively higher client rank, setting the matrix value of the agent / client pair to a value successively higher by a power of 2 N-1 Executing an operation including setting the matrix value of each agent / client pair in the bottom row to a value of 1, and for each agent / client pair having a successively higher client rank, setting the matrix value of the agent / client pair to a value successively higher by a power of 2, where N is the number of rows in which the agent / client pairs are arranged within the value matrix and the matrix value of the bottom row is 1, the computer program according to any one of clauses 13 to 15.

[0129] Clause 17 Based on the agent / client pairs in the high-value pair path that are emergency agent / client pairs and based on the agent / client pairs in the high-value pair path that are risky agent / client pairs, reducing the initial path value of the candidate pair path For each agent / client pair in the candidate pair path that is an emergency agent / client pair, reducing the initial path value of the high-value pair path by the matrix value of the emergency agent / client; and For each agent / client pair in the candidate path that is a risky agent / client pair, reducing the initial path value of the high-value pair path by a risk value based on the participation ability parameter of the risky agent / client pair The computer program according to any one of clauses 13 to 16, including the above.

[0130] Clause 18 Deriving the at least one participation decision based on the best path Based on the best path having an emergency agent / pair and in response to the participation ability of the emergency agent / pair being within a decision threshold or the start time being within a time threshold, determining a first action to be performed by the agent of the emergency agent / client pair Determining a second action to be performed by the agent of the agent / client pair in the best path in response to the agent of the agent / client pair in the best path that is not an emergency agent / client pair, wherein the second action to be performed by the agent of the agent / client pair in the best path is to stay on course comprising transmitting the at least one participation decision to the agent in the best path in response to determining the first action, transmitting the first action to the agent of the emergency agent / client pair transmitting the second action of staying on course to the agent of the agent / client pair in the best path that is not an emergency agent / client pair A computer program according to any one of clauses 10 to 17, comprising

[0131] Clause 19 An apparatus configured to generate a set of adjusted participation parameters by a processor in a multi-agent participation decision system, comprising at least one processor, and a memory communicatively coupled to the processor wherein the memory includes instructions executable by the processor, and when the instructions are executed by the processor, the processor is caused to determine agent / client pairs in at least one priority level based on the number of clients and the number of agents, each agent / client pair having an agent, a client, and a client rank of the client ​Determining a matrix value for each agent / client pair based on the client rank of the client of the agent / client pair; Determining a participation ability parameter for each agent / client pair; Setting the matrix value for the value of each rejected agent / ranked client to a zero value, where the rejected agent / client pair has a participation ability margin less than a predetermined threshold; Evaluating a plurality of possible participation options, where each participation option is a pair path from the agent / client pair with the highest client rank to the agent / client pair with the lowest client rank, and each pair path has an agent / client pair for each client in the at least one priority level; Calculating an initial path value for each pair path of the participation option; Determining a candidate pair path by determining the pair path with the highest initial path value, where the pair path with the highest initial path value is the candidate pair path; Reducing the initial path value of the candidate pair path for each candidate pair path based on the agent / client pair in the candidate pair path that is an emergency agent / client pair and based on the agent / client pair in the candidate pair path that is a risky agent / client pair, to derive a final path value for the candidate pair path; Determining the best path based on the final path value of each candidate pair path; Deriving at least one participation decision based on the best path; Sending the at least one participation decision to the agent within the best path An apparatus for performing operations including.

[0132] Article 20 Deriving the at least one engagement decision based on the best path comprises: Determining a first action for the agent of the emergency agent / client pair to perform in response to the best path having an emergency agent / pair and the engagement ability of the emergency agent / pair being within a decision threshold or the start time being within a time threshold; Determining a second action for the agent of the agent / client pair in the best path that is not an emergency agent / client pair, the second action for the agent of the agent / client pair in the best path being to stay on course; including: Sending the at least one engagement decision to the agent in the best path; In response to determining the first action, sending the first action to the agent of the emergency agent / client pair; Sending the second action of staying on course to the agent of the agent / client pair in the best path that is not an emergency agent / client pair; The apparatus according to clause 19, comprising:

[0133] As described above, the MAEDS described herein provides timely adjusted client engagement decisions, which is advantageous in preserving agent engagement capabilities, engaging with the same number of high-value clients as the number of agents (guaranteeing one-to-one service provision to the most valuable clients), and being flexible to changes in the client's environment. The dynamics of the client environment can include the progress of client arrivals, the recognition of client value, opportunities for engagement, management directives, agent resources, interference with engagement due to past service decisions, changes in the location and status of agents and clients, and the urgency of engagement.

[0134] The MAEDS described herein addresses the obstacles described herein. In particular, MAEDS addresses the difficulty of making multiple agent engagement decisions. This includes sorting out overlapping areas of client engagement capabilities, maintaining a balance between the ability to serve the highest value clients and the limits of agent resources, and mobilizing all agents so that the same number of agents engage with the same number of high-value clients. Additionally, MAEDS addresses the difficulty of making multiple agent engagement decisions that are adjusted in a timely manner. As the client value environment changes over time and the ability to recognize client value also changes, it becomes difficult to make decisions in a timely manner. MAEDS also addresses the difficulty of time and resource management for automatically adjusting to new change drivers as engagement evolves. These change drivers can include the arrival of new clients, unexpected obstacles and client activities, and unexpected agent health states and resource capabilities. Furthermore, MAEDS addresses the difficulty of avoiding hasty judgments that can lead to waste of agent engagement capabilities and a reduction in engagement options, resulting in a decline in client service.

[0135] Generally, all terms used in this specification shall be construed in accordance with their ordinary meaning in the relevant technical field, unless a different meaning is clearly imparted and / or suggested in the context in which the term is used. All references to an element, apparatus, component, means, step, etc. shall be construed openly as referring to at least one instance of that element, apparatus, component, means, step, etc., unless expressly stated otherwise. The steps of any method disclosed herein need not be performed in the exact order disclosed, unless it is clearly stated that one step follows or precedes another, and / or unless it is implicitly indicated that one step must follow or precede another. Any feature of any embodiment disclosed herein may, where appropriate, be applied to any other embodiment. Similarly, any advantage of any embodiment may be applied to any other embodiment, and vice versa. Other objects, features, and advantages of the disclosed embodiments will become apparent from the following description.

[0136] In the above description of various embodiments of the concept of the present invention, it should be understood that the technical terms used herein are for the purpose of explaining a particular embodiment and are not intended to limit the concept of the present invention. Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by those skilled in the technical field to which the concept of the present invention pertains. Furthermore, terms defined as in commonly used dictionaries shall be construed to have a meaning consistent with the meaning in the context of this specification and the relevant fields, and shall not be construed in an idealized or overly formal sense unless clearly defined herein.

[0137] If an element is described as "connected", "coupled", "responsive", or having a variant of these operations to another element, it may be directly connected, coupled, or responsive to the other element, or there may be intervening elements. In contrast, if an element is described as "directly connected", "directly coupled", "directly responsive", or having a variant of these operations to another element, there are no intervening elements. Similar numbers refer to similar elements throughout. Further, as used herein, "coupling", "connection", "response", or variants of these operations may include wireless coupling, connection, or response. As used herein, the singular forms with "a", "an", and "the" are intended to include the plural unless the context clearly dictates otherwise. Well-known functions or configurations may not be described for brevity and / or clarity. The term "and / or" includes any and all combinations of one or more of the associated listed items.

[0138] The terms first, second, third, etc. may be used herein to describe various elements / operations, but it should be understood that these terms should not be used to limit these elements / operations. These terms are only used to distinguish one element / operation from another. Thus, the first element / operation in some embodiments may be referred to as the second element / operation in other embodiments, so long as it does not depart from the teachings of the concept of the invention. The same reference numbers or the same reference designators indicate the same or similar elements throughout the specification.

[0139] As used herein, the expressions “comprises,” “comprised,” “comprising,” “includes,” “included,” “including,” “has,” “had,” “having,” or variations thereof are open-ended phrases and mean including one or more of the recited functions, integers, elements, steps, components, or functions, but do not exclude the presence or addition of one or more other functions, integers, elements, steps, components, functions, or groups thereof. Further, as used herein, the common abbreviation “e.g.” (for example), derived from the Latin “exempli gratia,” may be used to introduce or identify a general example of an item already described and is not intended to limit such item. The common abbreviation “i.e.” (that is), derived from the Latin “id est,” may be used to specify a particular item from a more general recitation.

[0140] In this specification, embodiments are described with reference to block diagrams and / or flowcharts of methods, apparatus (systems and / or devices), and / or computer program products implemented by computers. It will be understood that the blocks of the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by computer program instructions executed by one or more computer circuits. These computer program instructions may be provided to the processor circuits of general purpose computer circuits, special purpose computer circuits, and / or other programmable data processing circuits for constructing machines, and the instructions executed via the processor of the computer and / or other programmable data processing apparatus transform and control transistors, values stored in memory locations, and other hardware components within such circuits to implement the functions / acts specified in one or more blocks of the block diagrams and / or flowcharts, thereby constructing means (functionality) and / or structure for implementing the functions / acts specified in one or more blocks of the block diagrams and / or flowcharts.

[0141] Furthermore, these computer program instructions can also be stored in a tangible computer-readable medium that can direct a computer or other programmable data processing apparatus to function in a particular manner, whereby the instructions stored in the computer-readable medium, which include instructions for performing the functions / acts specified in one or more blocks of the block diagram and / or flowchart, produce a manufactured item. Accordingly, embodiments of the concepts of the present invention may be embodied in hardware and / or software (including firmware, resident software, microcode, etc.) executed by a processor such as a digital signal processor, and these may collectively be referred to as "circuitry," "module," or variations thereof.

[0142] It should further be noted that in some alternative implementations, the functions / acts noted in the blocks may be executed out of the order noted in the flowchart. For example, two blocks shown in succession may in fact be executed substantially concurrently or may sometimes be executed in the reverse order depending on the functions / acts involved. Further, the functionality of a given block of the flowchart and / or block diagram may be split into multiple blocks and / or the functionality of two or more blocks of the flowchart and / or block diagram may be at least partially integrated. Finally, other blocks may be added / inserted between the blocks shown and / or blocks / acts may be omitted, as long as this does not deviate from the scope of the concepts of the present invention. Additionally, although some of the figures include arrows on communication paths to indicate the primary direction of communication, it should be understood that communication may occur in the opposite direction to that shown by the arrows.

[0143] Many variations and modifications can be made to the embodiments without substantially departing from the principles of the concept of the present invention. All such variations and modifications are intended to be included within the scope of the concept of the present invention. Accordingly, the disclosed subject matter above should be regarded as illustrative and not restrictive, and the embodiments are intended to include all such modifications, improvements, and other embodiments that are within the spirit and scope of the concept of the present invention. Accordingly, to the maximum extent permitted by law, the scope of the concept of the present invention should be determined by the broadest permissible interpretation of the present disclosure, including the embodiments and their equivalents, and should not be restricted or limited by the foregoing detailed description.

Claims

1. A method for generating a set of adjusted participation parameters by a processor in a multi-agent participation decision system, comprising: determining agent / client pairs in at least one priority level based on the number of clients and the number of agents (701), each agent / client pair having an agent, a client, and a client rank of the client (701); determining a matrix value for each agent / client pair based on the client rank of the client in the agent / client pair (703); determining a participation ability parameter for each agent / client pair (705); setting the matrix value for each rejected agent / ranked client value to a zero value (707), where the rejected agent / client pair has a participation ability margin less than a predetermined threshold (707); evaluating a plurality of possible participation options (709), each participation option being a pair path from the agent / client pair with the highest client rank to the agent / client pair with the lowest client rank, each pair path having an agent / client pair for each client in the at least one priority level (709); calculating an initial path value for each pair path of the participation option for each participation option (711); determining a candidate pair path by determining the pair path having the highest initial path value (713), where the pair path having the highest initial path value is the candidate pair path (713); To derive a final path value for the candidate pair path, for each candidate pair path, based on the agent / client pair in the candidate pair path that is an emergency agent / client pair and based on the agent / client pair in the candidate pair path that is a risky agent / client pair, reducing the initial path value of the candidate pair path (715); Determining a best path based on the final path values of each candidate pair path (717); Deriving at least one participation decision based on the best path (719); Transmitting the at least one participation decision towards the agent within the best path (721); A method comprising. **Claim 2** In response to the participation ability parameter of the agent / client pair being within a range of a first threshold and a second threshold, the agent / client pair determines that it is an emergency agent / client pair (1101); In response to the participation ability parameter of the agent / client pair being less than a risk threshold, the agent / client pair determines that it is a risky agent / client pair (1103); The method according to claim 1, further comprising determining an agent / client pair that is an emergency agent / client pair and an agent / client pair that is a risky agent / client pair. **Claim 3** Determining agent / client pairs in the at least one priority level based on the number of clients and the number of agents, dynamically obtaining, from a status processor, the number of clients, the number of agents, and for each client, the client rank of the client (801); determining the at least one priority level based on the number of clients and the number of agents (803); grouping some agent / client pairs into the at least one priority level according to the client rank of the clients in the agent / client pairs (805); The method according to claim 1 or 2, comprising. **Claim 4** Deriving the value matrix for the at least one priority level (901) based on the number of agents and the number of clients in the at least one priority level, the value matrix having a number of rows less than or equal to the number of agents for which a decision is required and less than or equal to the number of clients, the method according to any one of claims 1 to 3, further comprising deriving the value matrix for the at least one priority level (901).

5. The method according to claim 4, further comprising enabling more than one agent to be involved with a client by adding lower ranked rows to the value matrix for the client (903).

6. Each engagement option is a pair path from the agent / client pair of the top row of the value matrix to the agent / client pairs of each intermediate row of the value matrix and to the agent / client pair of the bottom row of the value matrix, each agent and client of each agent / client in the pair path being different from the other agents and clients of the agent / client pair of the pair path, the method according to claim 4 or 5.

7. Setting the matrix value of each agent / client pair of the bottom row of the value matrix to a value of 1 (1001); Setting the matrix values of each agent / client pair having successively higher client ranks to successively higher values by a power of 2 (1003); Ｎ－１ and further comprising, where N is the number of rows in which the agent / client pair is arranged within the value matrix and the matrix value of the bottom row is 1, the method according to any one of claims 4 to 6.

8. Reducing the initial path value of the candidate pair path based on the agent / client pairs in the high value pair path that are emergency agent / client pairs and based on the agent / client pairs in the high value pair path that are risky agent / client pairs, for each agent / client pair in the candidate pair path that is an emergency agent / client pair, reducing the initial path value of the candidate pair path by only the matrix value of the emergency agent / client (1201). For each agent / client pair in the candidate pair path that is a risky agent / client pair, reducing the initial path value of the candidate pair path by only the risk value based on the participation ability parameter of the risky agent / client pair (1203); The method according to claim 7, comprising:

9. Deriving the at least one participation decision based on the best path comprises: In response to the best path having an emergency agent / pair and the participation ability parameter of the emergency agent / pair being within a decision threshold or the start time being within a time threshold, determining a first action to be performed by the agent of the emergency agent / client pair (1301); Determining a second action to be performed by the agent of the agent / client pair in the best path in response to the agent of the agent / client pair in the best path that is not an emergency agent / client pair, wherein the second action to be performed by the agent of the agent / client pair in the best path is to stay on course (1305); comprising: Sending the at least one participation decision to the agent in the best path; In response to determining the first action, sending the first action to the agent of the emergency agent / client pair (1303); Sending the second action of staying on course to the agent of the agent / client pair in the best path that is not an emergency agent / client pair (1307); The method according to any one of claims 1 to 8, comprising:

10. A computer program having a non-transitory computer-readable medium with computer-executable instructions, which, when executed by a processor included in a device, cause the device to: Determining an agent / client pair in at least one priority level based on the number of clients and the number of agents (701), wherein each agent / client pair has an agent, a client, and a client rank of the client, determining the agent / client pair (701); Determining a matrix value for each agent / client pair based on the client rank of the client of the agent / client pair (703); Determining a participation ability parameter for each agent / client pair (705); Setting the matrix value for each value of each rejected agent / ranked client to a zero value (707), wherein the rejected agent / client pair has a participation ability margin less than a predetermined threshold, setting the matrix value to a zero value (707); Evaluating a plurality of possible participation options (709), wherein each participation option is a pair path from an agent / client pair having the highest client rank to an agent / client pair having the lowest client rank, and each pair path has an agent / client pair for each client in the at least one priority level, evaluating the plurality of possible participation options (709); Calculating an initial path value for each pair path of the participation option for each participation option (711); Determining a candidate pair path by determining a pair path having the highest initial path value (713), wherein the pair path having the highest initial path value is the candidate pair path, determining the candidate pair path (713); Reducing the initial path value of each candidate pair path based on the agent / client pair in the candidate pair path that is an urgent agent / client pair and based on the agent / client pair in the candidate pair path that is a risky agent / client pair to derive a final path value for the candidate pair path (715); Determining an optimal path based on the final path value of each candidate pair path (717); Deriving at least one engagement decision based on the optimal path (719); Transmitting the at least one engagement decision to an agent within the optimal path (721); A computer program for causing an operation including the above to be executed. **Claim 11** The non-transitory computer-readable medium includes further computer-executable instructions that, when executed, cause the device to Determine that an agent / client pair is an emergency agent / client pair in response to the engagement capability parameter of the agent / client pair being within a range of a first threshold and a second threshold (1101); Determine that an agent / client pair is a risky agent / client pair in response to the engagement capability parameter of the agent / client pair being less than a risk threshold (1103); Determining an agent / client pair that is an emergency agent / client pair and an agent / client pair that is a risky agent / client pair The computer program according to claim 10, for causing an operation including the above to be executed. **Claim 12** Determining an agent / client pair in at least one priority level based on the number of clients and the number of agents is Dynamically obtaining, from a status processor, the number of clients, the number of agents, and for each client, the client rank of the client (801); Determining the at least one priority level based on the number of clients and the number of agents (803); Grouping some agent / client pairs into at least one priority level according to the client rank of the client in the agent / client pair (805); The computer program according to claim 10 or 11, including the above. **Claim 13** The non-transitory computer-readable medium includes further computer-executable instructions that, when executed, cause the device to Deriving a value matrix for the at least one priority level based on the number of agents and the number of clients in the at least one priority level (901), wherein the value matrix has a number of rows less than or equal to the number of agents for which a decision is required and less than or equal to the number of clients. Deriving a value matrix for the at least one priority level (901). A computer program according to any one of claims 10 to 12, which causes an operation including this to be executed. **Claim 14** The non-transitory computer-readable medium comprises further computer-executable instructions that, when executed, cause the device to perform operations including enabling more than one agent to participate in the client by adding lower-ranked rows to the value matrix for the client (903). A computer program according to claim 13. **Claim 15** Each participation option is a pair path from the agent / client pair in the top row of the value matrix to the agent / client pairs in each middle row of the value matrix and to the agent / client pair in the bottom row of the value matrix, and each agent and client in each agent / client in the pair path is different from the other agents and clients in the agent / client pair of the pair path. A computer program according to claim 14. **Claim 16** The non-transitory computer-readable medium comprises further computer-executable instructions that, when executed, cause the device to set the matrix value of each agent / client pair in the bottom row of the value matrix to a value of 1 (1001); Setting the matrix values of each agent / client pair having successively higher client ranks to successively higher values by a power of 2 (1003) and Ｎ－１ ​ perform operations including this; where N is the number of rows in which the agent / client pair is arranged within the value matrix, and the matrix value of the bottom row is 1. A computer program according to any one of claims 13 to 15. **Claim 17** Based on the agent / client pairs in the high-value pair path that are emergency agent / client pairs and based on the agent / client pairs in the high-value pair path that are risky agent / client pairs, reducing the initial path value of the candidate pair path is For each agent / client pair in the candidate pair path that is an emergency agent / client pair, reducing the initial path value of the candidate pair path by only the matrix value of the emergency agent / client (1201); For each agent / client pair in the candidate pair path that is a risky agent / client pair, reducing the initial path value of the candidate pair path by only the risk value based on the participation ability parameter of the risky agent / client pair (1203) The computer program according to claim 16, comprising:

18. Deriving the at least one participation decision based on the optimal path is In response to the optimal path having an emergency agent / pair and the participation ability parameter of the emergency agent / pair being within a decision threshold or the start time being within a time threshold, determining the first action to be performed by the agent of the emergency agent / client pair (1301); Determining the second action to be performed by the agent of the agent / client pair in the optimal path in response to the agent of the agent / client pair in the optimal path that is not an emergency agent / client pair, wherein the second action to be performed by the agent of the agent / client pair in the optimal path is to stay on course (1305) including Sending the at least one participation decision to the agent in the optimal path is In response to determining the first action, sending the first action to the agent of the emergency agent / client pair (1303); Transmitting the second action of staying on the course to the agent of the agent / client pair on the best path, which is not an emergency agent / client pair (1307); and A computer program according to any one of claims 10 to 17, comprising: **Claim 19** An apparatus configured to generate a set of adjusted participation parameters by a processor in a multi-agent participation determination system, comprising: At least one processor, and A memory communicatively coupled to the processor wherein the memory includes instructions executable by the processor, the instructions causing the processor to: Determine an agent / client pair in at least one priority level based on the number of clients and the number of agents (701), each agent / client pair having an agent, a client, and a client rank of the client; (701); Determine a matrix value for each agent / client pair based on the client rank of the client of the agent / client pair (703); Determine a participation ability parameter for each agent / client pair (705); Set the matrix value to a zero value for each discarded agent / ranked client value, where the discarded agent / client pair has a participation ability margin below a predetermined threshold; (707); Evaluate a plurality of possible participation options (709), each participation option being a pair path from an agent / client pair having the highest client rank to an agent / client pair having the lowest client rank, each pair path having an agent / client pair for each client in the at least one priority level; (709); For each participation option, calculate an initial path value for each pair path of the participation option (711); ​ Determining candidate pair paths by determining a pair path having the highest initial path value (713), wherein the pair path having the highest initial path value is the candidate pair path; determining candidate pair paths (713); For each candidate pair path, reducing the initial path value of the candidate pair path based on the agent / client pair in the candidate pair path that is an emergency agent / client pair and based on the agent / client pair in the candidate pair path that is a risky agent / client pair to derive a final path value for the candidate pair path (715); Determining the best path based on the final path value of each candidate pair path (717); Deriving at least one involvement decision based on the best path (719); Transmitting the at least one involvement decision to the agent within the best path (721); An apparatus for performing operations including. **Claim 20** Deriving the at least one involvement decision based on the best path includes: In response to the best path having an emergency agent / pair and the involvement capability parameter of the emergency agent / pair being within a decision threshold or the start time being within a time threshold, determining a first action to be performed by the agent of the emergency agent / client pair (1301); Determining a second action to be performed by the agent of the agent / client pair in the best path in response to the agent of the agent / client pair in the best path that is not an emergency agent / client pair, wherein the second action to be performed by the agent of the agent / client pair in the best path is to stay on course (1305); determining the second action (1305) Including; Transmitting the at least one involvement decision to the agent in the best path includes: In response to determining the first action, transmitting the first action to the agent of the emergency agent / client pair (1303); Transmitting the second action of staying on the course to the agent of the agent / client pair on the best path, which is not an emergency agent / client pair (1307); and The apparatus according to claim 19, comprising:

Images