Edge computing resource allocation optimization method and system based on intelligent agent negotiation

Abstract

The edge computing resource allocation optimization method and system based on agent negotiation provided by the present invention are applied to consumer agents and relate to the field of edge computing technology. The present invention broadcasts a negotiation topic request for resource allocation to multiple resource provider agents in the communication area, i.e., a quotation; receives a resource allocation solution provided by the provider agent, i.e., a counter-offer; evaluates the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the negotiation is successful, and the service quality requirements are determined with the selected provider agent to achieve resource allocation; otherwise, a new behavior state is obtained through concession calculation in combination with the satisfaction level; a feasible solution set is generated according to the new behavior state; an expected solution is obtained according to the feasible solution set and the counter-offer, and a new quotation is generated to conduct a new round of negotiation topic requests until the negotiation termination condition is met. The present invention can dynamically adjust resource allocation and solve the problem of low resource allocation efficiency in edge computing.

Description

Technical Field

[0001] The present invention relates to the field of edge computing technology, and in particular to an edge computing resource allocation optimization method based on intelligent agent negotiation and a system thereof. Background Art

[0002] The traditional cloud computing model faces problems such as network latency and privacy security due to its centralized processing characteristics, and it is difficult to meet the needs of low latency, high quality of experience (QoE) and diversified application services. Therefore, edge computing came into being, which provides computing, storage and processing capabilities at the edge of the network to reduce latency, increase data processing speed, and reduce the burden on the central cloud.

[0003] However, edge computing also faces challenges of limited resources and highly dynamic and uncertain environments. Frequent joining and leaving of users, as well as rapid changes in task requirements and resource availability, make it difficult for existing resource allocation algorithms to effectively cope with this dynamic and complex environment. They are inefficient and lack flexibility when dealing with multi-issue negotiations, resulting in inefficient resource allocation and failure to meet end-user quality of service (QoS) requirements. These factors lead to huge challenges in resource allocation and task offloading in edge computing, requiring a more intelligent and flexible approach to cope with the ever-changing environment.

[0004] In view of this, the applicant filed this application after studying the existing technology. Summary of the invention

[0005] The present invention aims to provide an edge computing resource allocation optimization method and system based on intelligent agent negotiation to solve the shortcomings of existing methods such as low resource allocation efficiency and inability to meet the end users' requirements for service quality.

[0006] In order to solve the above technical problems, the present invention is implemented through the following technical solutions:

[0007] An edge computing resource allocation optimization method based on agent negotiation, applied to consumer agent CA, includes the following steps:

[0008] S1, broadcasts its resource allocation negotiation topic request, i.e., quotation, to multiple provider agents PA in the communication area;

[0009] S2, receiving the resource allocation solution provided by PA, i.e., counter-offer;

[0010] S3, evaluate the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the selected PA to achieve resource allocation; otherwise, combined with the satisfaction level, a new behavior state is obtained through concession calculation; wherein the new behavior state is the satisfaction of the agent with the current resource allocation solution;

[0011] S4, generating a feasible solution set according to the new behavior state;

[0012] S5, selecting the optimal solution according to the feasible solution set and the counter-offer to obtain the expected solution;

[0013] S6, generating a new quotation based on the expected solution to request a new round of negotiation topics until the negotiation termination condition is met.

[0014] Preferably, when evaluating the resource allocation solution, an aggregate utility value is calculated through a utility function to express the agent's preference for the resource allocation solution; wherein the utility function is expressed as:

[0015] ;

[0016] in, Indicates the solution S The aggregate utility value of each agent; Indicates the solution S The fuzzy membership function of the negotiation topics; Indicates the number of negotiation topics being negotiated; Indicates The weight factor of each negotiation topic.

[0017] Preferably, the satisfaction level is reflected by constraints on self-satisfaction, opponent's response information, time limit and market factors; wherein,

[0018] Self-satisfaction It means that the definition is as follows:

[0019] ;

[0020] in, Indicates potential solutions; Indicates the The aggregate utility value of each agent;

[0021] The opponent's response information is through the quotation Counter-offer with the opponent The degree of response between It is defined as:

[0022] ;

[0023] in, Indicates the degree of response; Indicates the nth round of bidding; represents the opponent's nth round of counter-offer; n represents the current negotiation round; represents the n-1th round of bidding, i.e. the previous round of bidding; D represents the distance between the bid A and the counter-offer B, which is calculated by the following formula;

[0024] ;

[0025] in, Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; represents the distance measure between two fuzzy sets, where the fuzzy sets represent the probability distribution of offer A and counter-offer B in each negotiation topic of the agent;

[0026] The time limit constraint is used to ensure that the negotiation process is carried out within an acceptable time frame and is defined as follows:

[0027] ;

[0028] in, Indicates the negotiation time of the current negotiation round; Indicates the negotiation deadline of the current negotiation round; represents a constant, The initial value of ; represents the slope control parameter, when When , the time curve is steeper, and the concession becomes larger as n increases;

[0029] Market factors evaluate the ratio of market supply and demand by considering the difference between the ratio of the offer and the opponent's concession. If the concession value of the current negotiation round is less than the average concession value, the demand is less than the supply. Market factors is defined as follows:

[0030] ;

[0031] in, represents the distance between offer A and counter-offer B in the nth round of negotiation; Indicates the average distance between the offer and the counteroffer in the previous negotiation rounds at the nth negotiation round.

[0032] Preferably, the new behavior state is obtained by concession calculation, specifically:

[0033] Based on the current negotiation cycle's own satisfaction , responsiveness , Negotiation time and market factors , calculate the agent's concession value , the formula is as follows:

[0034] ;

[0035] in, , , , According to their own satisfaction , Response level , Negotiation time and market factors The concession conditions obtained by the constraints;

[0036] Based on the calculated handicap value , get the new behavior state :

[0037] ;

[0038] in, Indicates the current behavior status.

[0039] Preferably, the formula of the feasible solution set P is as follows:

[0040] ;

[0041] in, Indicates a solution; Indicates a new behavior state; Indicates the new intention obtained according to the new behavior state; Indicates A fuzzy constraint network; Indicates the solution S The aggregate utility value of each agent; Indicates the current behavior status.

[0042] Preferably, the desired solution The formula is:

[0043] ;

[0044] in, represents the solution; B represents the counter-offer; P represents the feasible solution set; ) represents the metric function, and its calculation formula is:

[0045] ;

[0046] in, Indicates the number of negotiation topics being negotiated; Indicates the solution S The fuzzy membership function of the negotiation topics; represents the weight factor associated with solution satisfaction; Respectively represent the similarity weight factors with solution satisfaction; Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; It represents the distance measure between two fuzzy sets, where the fuzzy sets represent the possibility distribution of offer A and counteroffer B in each negotiation topic.

[0047] Preferably, the new quotation is: generating a new quotation element for each object of the current negotiation topic, and the new quotation is composed of the new quotation elements, which is expressed as:

[0048] ;

[0049] ;

[0050] in, indicating said new offer; Represents the new quote element of the qth object, that is, the kth agent in the space The marginal likelihood distribution in ; Indicates the total number of objects of the current negotiation topic; is the expected solution;

[0051] The kth agent of the qth object is in the space Cylinder expansion in ; represents the space of the qth object and the kth agent; Proj represents the space projection function.

[0052] Preferably, the negotiation termination condition includes negotiation success or negotiation failure; wherein,

[0053] Negotiation success is defined as: the aggregate utility value of the current counter-offer is greater than the new behavior state, or the expected solution is The aggregate utility value of is greater than the new behavior state, expressed as:

[0054] ;

[0055] in, Desire solution Middle The aggregate utility value of each agent; Indicates a new behavior state; represents the aggregate utility value of the current counter-offer B;

[0056] Negotiation failure is defined as: no new solution can be found or negotiation resources are exhausted.

[0057] Preferably, when providing a resource allocation solution, the PA evaluates the CA's quotation. If the evaluation result is to accept the quotation, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the current CA to allocate resources to it. Otherwise, based on the satisfaction level of the PA, a feasible solution set is generated through concession calculation; the expected solution is obtained based on the feasible solution; based on the expected solution, a resource allocation solution, i.e., a counter-offer, is generated until the negotiation termination conditions are met.

[0058] The present invention also provides an edge computing resource allocation optimization system based on agent negotiation, comprising: multiple consumer agents CA and multiple provider agents PA; multiple CAs initiate negotiation requests by broadcasting resource quotes to multiple PAs in a communication area; CAs and PAs continuously exchange quotes and counter-offers until the negotiation is terminated; wherein,

[0059] The consumer agent CA is used to: broadcast its resource allocation negotiation topic request, i.e., quotation, to multiple provider agents PA in the communication area; receive the resource allocation solution provided by PA, i.e., counter-offer; evaluate the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the selected PA to achieve resource allocation; otherwise, combined with the satisfaction level, a new behavior state is obtained through concession calculation; wherein the new behavior state is the satisfaction of the agent with the current resource allocation solution; based on the new behavior state, a feasible solution set is generated; based on the feasible solution set, the optimal solution is selected to obtain the expected solution; based on the expected solution, a new quotation is generated to conduct a new round of negotiation topics until the negotiation termination condition is met;

[0060] The provider intelligent agent PA is used to: receive the quotation of CA for evaluation; if the evaluation result is to accept the quotation, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the current CA to allocate resources for it; otherwise, combined with the satisfaction level, a feasible solution set is generated through concession calculation; the expected solution is obtained based on the feasible solution set; based on the expected solution, a resource allocation solution, i.e., a counter-offer, is generated until the negotiation termination condition is met; based on the expected solution, a resource allocation solution, i.e., a counter-offer, is generated until the negotiation termination condition is met.

[0061] The present invention also provides an edge computing resource allocation optimization device based on intelligent agent negotiation, comprising a processor and a memory, wherein a computer program is stored in the memory, and the computer program can be executed by the processor to implement an edge computing resource allocation optimization method based on intelligent agent negotiation as described above.

[0062] The present invention also provides a computer-readable storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a device where the computer-readable storage medium is located, an edge computing resource allocation optimization method based on agent negotiation as described above is implemented.

[0063] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0064] The present invention solves the problem of edge computing resource allocation by constructing a unified multilateral and multi-issue negotiation framework to handle the negotiation problems between users and edge devices, thereby improving the flexibility and efficiency of negotiation. The present invention uses fuzzy constraint functions to model resource allocation negotiation topics, so as to interpret the negotiation information of resource allocation in a unified, flexible and comprehensive manner, and solve the problem of information uncertainty and inaccuracy in the process of edge computing resource allocation.

[0065] The present invention transforms the negotiation problem into a distributed fuzzy constraint satisfaction problem, and through the mutual association of constraints between intelligent agents, continuously relaxes the constraint requirements (i.e., price concession calculation) during the negotiation process, generates a solution that satisfies all constraints, generates a new quotation for a new round of negotiation, and finally satisfies the negotiation termination conditions. The overall negotiation process includes five steps: proposal generation, proposal evaluation, calculation of feasible solutions, quotation generation, and negotiation termination.

[0066] The intelligent negotiation technology of the present invention can dynamically adjust the resource allocation strategy according to the real-time diversified user needs and changing resource availability to adapt to the dynamic and uncertain environment, so as to optimize resource allocation, meet the service quality requirements of diversified users, improve resource utilization and cost-effectiveness, and enhance the adaptability and reliability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0068] Figure 1 A flowchart of an edge computing resource allocation optimization method based on agent negotiation provided in Example 1.

[0069] Figure 2 A structural diagram of an edge computing resource allocation optimization method based on agent negotiation provided in Example 1.

[0070] Figure 3 A structural diagram of an edge computing resource allocation optimization system based on agent negotiation provided in Example 1.

[0071] Figure 4 This is a structural diagram of a distributed fuzzy constraint network based on agent negotiation provided in Example 1.

[0072] Figure 5 This is a curve comparison chart of the average aggregate utility values ​​successfully negotiated between multiple PAs and multiple CAs under multiple edge computing algorithms provided in Example 1.

[0073] Figure 6 This is a curve comparison chart of the negotiation success ratios of multiple PAs and multiple CAs under multiple edge computing algorithms provided in Example 1.

[0074] Figure 7 This is a curve comparison chart of the CPU utilization of edge servers used for successful requests under multiple edge computing algorithms provided in Example 1.

[0075] Figure 8 This is a curve comparison chart of the standard deviation change of load under multiple edge computing algorithms provided in Example 1.

[0076] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. DETAILED DESCRIPTION

[0077] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0078] Embodiment 1

[0079] Embodiment 1 of the present invention provides an edge computing resource allocation optimization method based on intelligent agent negotiation, which can be implemented by an edge computing resource allocation optimization device based on intelligent agent negotiation (hereinafter referred to as the resource allocation device), and in particular, executed by one or more processors in the resource allocation device.

[0080] In this embodiment, the resource allocation device may be an electronic device equipped with a processor, which has a computer program of the edge computing resource allocation optimization method based on agent negotiation and can be executed, such as a computer, a smart phone, a smart tablet, a workstation, etc., which is not limited here.

[0081] In edge computing scenarios, intelligent negotiation technology can dynamically adjust resource allocation strategies based on real-time diversified user needs and changing resource availability to adapt to the dynamics and uncertainty of the environment, optimize resource allocation, improve resource utilization and cost-effectiveness, meet user personalized needs, and enhance system adaptability. The negotiation model of intelligent agents can be applied to multiple fields, such as:

[0082] Smart home: Agents can negotiate how to allocate household resources (such as electricity, network bandwidth, etc.) to achieve more efficient energy management and network optimization.

[0083] Intelligent Transportation: Intelligent agents can negotiate how to dispatch traffic flow, optimize signal light control, etc. to alleviate traffic congestion and improve road efficiency.

[0084] Industrial Automation: Intelligent agents can negotiate how to allocate production tasks, optimize production processes, etc. to improve production efficiency and product quality.

[0085] However, when using intelligent negotiation to allocate edge computing resources, we face the following key technical issues:

[0086] (1) Information sharing and uncertainty processing. In edge computing resource negotiation, the uncertainty and imprecision of user service demand information, as well as the incomplete sharing of some information (such as user budget and supplier resource cost), affect the final decision-making behavior and effect. Existing technologies lack effective methods to represent this negotiation information in a unified, flexible and comprehensive manner. Therefore, it is necessary to develop an algorithm that can handle information sharing limitations and uncertainties to improve decision quality and system robustness.

[0087] (2) Complexity of resource allocation model and algorithm generalization ability. Edge computing resource allocation models face complexity challenges, which not only involve the coordination of multiple constraints such as cost budget, quality of service (QoS) requirements and resource capacity constraints, but also require algorithms to have good generalization capabilities to adapt to different environments and conditions. Existing auction and game theory methods are inefficient and lack flexibility when dealing with multi-issue negotiations. At the same time, the models trained by these methods in specific environments are often difficult to generalize to other environments, limiting the applicability and effectiveness of the algorithms.

[0088] (3) Challenges of dynamic resource allocation and adaptation to environmental dynamics. Faced with the randomness and dynamic characteristics of user tasks, existing resource allocation algorithms are difficult to adapt to unknown changes in the environment and long-term benefits, resulting in inefficient resource allocation and resource waste. This requires us to develop a resource allocation algorithm that can cope with the randomness and dynamics of user tasks, as well as environmental dynamics (including node changes), to achieve efficient resource utilization and optimal allocation.

[0089] like Figure 1-Figure 2 As shown, the present invention proposes an edge computing resource allocation optimization method based on agent negotiation, which is applied to consumer agent CA and includes steps S1 to S6.

[0090] S1, broadcasts its resource allocation negotiation topic request, i.e., quotation, to multiple provider agents PA in the communication area;

[0091] S2, receives the resource allocation solution provided by PA, i.e., the counter-offer.

[0092] like Figure 3 As shown in the figure, there are two types of agents: Consumer Agent (CA) and Provider Agent (PA). CA broadcasts its service request and turns to the negotiation queue. Each request of CA will be negotiated with multiple PAs. In each round of negotiation, CA submits a quotation, and then each PA counter-offers separately. CA will receive all the information of the counter-offer, while PA only knows the information provided by CA. If the negotiation is successful, CA will determine the service quality requirement SLA (Service Level Agreement) with the selected PA, and the selected PA will immediately allocate resources. This negotiation model is consistent with the transaction method in the real world. Buyers tend to negotiate with multiple sellers and decide the final seller after accessing the information of all sellers. This model uses a negotiation queue, and one representation of the queue negotiates the CA process.

[0093] Consumer Agent (CA): represents the consumer user of edge computing resources. Based on the needs of consumer preferences, it can automatically enter the service market of edge computing resources. For example, some consumers want low prices for resources, and some prefer low response times. These preferences represent specific service quality QoS service requests constructed by CA and restrict each other in CA.

[0094] Provider Agent (PA): represents the edge computing resource service provider. Generally speaking, PA hopes to gain more benefits in the market. PA adopts reasonable resource allocation to reduce costs. On the other hand, PA adopts competition strategy to meet more consumer requests.

[0095] Quality of Service Requirements SLA: SLA specifies the details of the functional and non-functional requirements of the consumer for the service, called Quality of Service (QoS) requirements. Negotiations between CA and PA finalize the SLA with specified prices and penalties.

[0096] like Figure 4 As shown in the figure, in the SLA negotiation process, the shared information between agents is often uncertain and imprecise, which will affect the decision-making behavior of the agents. In the face of this problem, fuzzy constraints can be used to model the negotiation topics to represent the negotiation information as flexibly and comprehensively as possible. Fuzzy constraints are not only used to define each agent's beliefs involving imprecise and semantic concepts, but can also be used to represent the relationship between agents.

[0097] Therefore, the SLA problem can be described as a distributed fuzzy constraint satisfaction problem (DFCSP), which is interconnected by constraints between agents, and a solution that satisfies all constraints is generated through negotiation to achieve a mutually satisfactory result. Each fuzzy constraint network FCN contains a set of objects, restrictions, and goals, and is controlled by agents. Therefore, the DFCSP problem can be solved by agent negotiation, where the fuzzy relations between agents represent all constraints in the distributed fuzzy constraint network (DFCN).

[0098] The distributed fuzzy constraint network DFCN can be defined as follows:

[0099] DFCN={U,X,C};

[0100] Among them, U is the domain of the entire DFCN;

[0101] is the set of all non-recursive objects in DFCN, and is the non-recursive object of the kth agent; N represents the number of agents.

[0102] is the set of all fuzzy constraints in DFCN, is a set of fuzzy constraints for the kth agent, involving A set of internal fuzzy constraints between objects in and a set of external fuzzy constraints between the kth agent.

[0103] A set of fuzzy constraint networks (FCN) can be defined as:

[0104] ;

[0105] , represents the set of fuzzy constraint networks; represents the number of fuzzy constraint networks; Indicates A fuzzy constraint network.

[0106] , indicating the A fuzzy constraint network, that is, the kth agent passes A set of external constraints is connected to the fuzzy constraint network of other sets;

[0107] in, represents the domain of the kth Agent. It represents a set of fuzzy constraints of the kth agent, including internal constraints and external constraints, which also corresponds to a set of restrictions. Therefore, to solve DFCSP by using SLA negotiation, it is necessary to meet the constraints through negotiation according to the intention of the agent and finally realize the SLA.

[0108] Each agent participating in the negotiation can correspond to a fuzzy constraint network, and the negotiation between agents corresponds to each object with constraints; the needs and preferences of the agents can also be expressed through fuzzy constraints.

[0109] Fuzzy Constraint Network The intention is expressed as ,yes The probability distribution of objects in is defined as follows:

[0110] ;

[0111] in, Indicates that the i-th fuzzy constraint is in the space The projection in , m represents the number of projections.

[0112] and As The fuzzy segmentation satisfies Acceptable threshold All internal and external constraints The set of solutions of . If , indicating that there is no feasible solution, Indicates that it is empty. The agent will adjust the threshold , and utilizes fuzzy constraint relaxation to reconfigure the scope of constraints, thereby generating new feasible solutions, thus moving all constraints in DFCN toward a common satisfactory solution. Therefore, solving DFSCP, a satisfactory global solution can be obtained.

[0113] S3, evaluate the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the selected PA to achieve resource allocation; otherwise, combined with the satisfaction level, a new behavior state is obtained through concession calculation; wherein, the new behavior state is the agent's satisfaction with the current resource allocation solution.

[0114] In this embodiment, multiple CAs start negotiation requests by proposing ideal resource quotations to corresponding multiple PAs, and then CAs and PAs continuously exchange quotations and counter-offers until the negotiation terminates.

[0115] Agent negotiation includes the following steps: solution evaluation, concession calculation, feasible solution generation, offer generation, and negotiation termination. First, the CA or PA evaluates whether the offer or counteroffer is acceptable for the solution. If the solution is unacceptable, the Agent (CA or PA) will determine the new behavior state by evaluating the internal behavior state, the opponent's response information, the negotiation time factor, and the market factor. Then, the Agent generates a set of feasible solutions based on the new behavior state, and packages them into an offer or counteroffer to the opponent Agent by searching for the optimal potential solution. Therefore, the Agent program can iteratively exchange new proposals until the termination condition (such as consensus or failure) is met.

[0116] When evaluating the resource allocation solution, the aggregate utility value (Aggregation Utility, AU) is calculated through the utility function to express the agent's preference for the resource allocation solution (i.e., offer or counteroffer) to decide whether to reach an agreement or make concessions. The utility function is expressed as:

[0117] ;

[0118] in, Indicates the solution S The aggregate utility value of each agent is used to determine whether the solution S meets the current intention. and the current behavior status . Behavioral state It is used to estimate whether the agent will accept the solution S under the current level of constraint satisfaction, or whether it needs to propose a new negotiation topic or make a counter-offer at a lower behavior state.

[0119] The concession calculation strategy will guide the agent's behavior by lowering the cutting threshold To adjust the intention , and generate a new proposal, thus moving the negotiation toward a common solution.

[0120] Indicates the solution S The fuzzy membership function can help agents evaluate the imprecise preferences of multiple issues. Indicates the number of negotiation topics being negotiated; Indicates The weight factor of each negotiation topic.

[0121] Furthermore, the satisfaction level is reflected by the constraints of self-satisfaction, opponent's response information, time limit and market factors; wherein,

[0122] Self-satisfaction It means that the definition is as follows:

[0123] ;

[0124] in, Indicates potential solutions; Indicates the The aggregate utility value of each agent;

[0125] The opponent's response information is through the quotation Counter-offer with the opponent The degree of response between It is defined as:

[0126] ;

[0127] in, Indicates the degree of response; Indicates the nth round of bidding; represents the opponent's nth round of counter-offer; n represents the current negotiation round; represents the n-1th round of bidding, i.e. the previous round of bidding; D represents the distance between the bid A and the counter-offer B, which is calculated by the following formula;

[0128] ;

[0129] in, Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; represents the distance measure between two fuzzy sets, where the fuzzy sets represent the probability distribution of offer A and counter-offer B in each negotiation topic of the agent;

[0130] The time limit constraint is used to ensure that the negotiation process is carried out within an acceptable time frame and is defined as follows:

[0131] ;

[0132] in, Indicates the negotiation time of the current negotiation round; Indicates the negotiation deadline of the current negotiation round; represents a constant, The initial value of ; Slope control parameter, when When , the time curve is steeper, and the concession becomes larger as n increases;

[0133] Agents cannot obtain the entire information of the dynamic market. Market factors evaluate the ratio of market supply and demand by considering the difference between the bid and the opponent's concession. If the concession value of the current negotiation round is less than the average concession value, the demand is less than the supply, and vice versa. is defined as follows:

[0134] ;

[0135] in, It represents the distance between offer A and counter-offer B in the nth round of negotiation; Indicates the average distance between the offer and the counteroffer in the previous negotiation rounds at the nth negotiation round.

[0136] The new behavior state is obtained by concession calculation, specifically:

[0137] Based on the current negotiation cycle's own satisfaction , responsiveness , Negotiation time and market factors , calculate the agent's concession value , the formula is as follows:

[0138] ;

[0139] in, , , , According to their own satisfaction , responsiveness , Negotiation time and market factors The concession conditions are obtained based on the constraints of . The concession strategy can determine how the agent can flexibly transfer its behavior state in dynamically changing resources. In particular, the opponent's behavior information with fuzzy membership function contains the opponent's preference information and can generate satisfactory solutions more effectively.

[0140] Based on the calculated handicap value , get the new behavior state :

[0141] ;

[0142] in, Indicates the current behavior status.

[0143] S4: Generate a feasible solution set based on the new behavior state.

[0144] The formula of the feasible solution set P is as follows:

[0145] ;

[0146] in, Indicates a solution; Indicates a new behavior state; Indicates the new intention obtained according to the new behavior state; Indicates the solution S The aggregate utility value of each agent; Indicates the current behavior status.

[0147] S5, selecting the optimal solution according to the feasible solution set and the counter-offer to obtain the expected solution.

[0148] The agent uses fuzzy constraints to specify preferences for negotiation topics, sorts feasible solutions, and obtains the expected solution. The formula is:

[0149] ;

[0150] in, represents the solution; B represents the counter-offer; P represents the feasible solution set; ) represents the metric function, and its calculation formula is:

[0151] ;

[0152] in, Indicates the number of negotiation topics being negotiated; Indicates the solution S The fuzzy membership function of the negotiation topics; represents the weight factor associated with solution satisfaction; Respectively represent the similarity weight factors with solution satisfaction; Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; It represents the distance measure between two fuzzy sets, where the fuzzy sets represent the possibility distribution of offer A and counteroffer B in each negotiation topic.

[0153] S6, generating a new quotation based on the expected solution to conduct a new round of negotiation topics until the negotiation termination conditions are met.

[0154] Get feasible solution set P and expected solution , you can generate new quotes on the topic's object set.

[0155] The new quotation is: by generating a new quotation element for each object of the current negotiation topic, the new quotation is composed of the new quotation elements, which is expressed as:

[0156] ;

[0157] ;

[0158] in, indicating said new offer; Represents the new quote element of the qth object, that is, the kth agent in the space The marginal likelihood distribution in ; Indicates the total number of objects of the current negotiation topic; is the expected solution;

[0159] The kth agent of the qth object is in the space Cylinder expansion in ; represents the space of the qth object and the kth agent; Proj represents the space projection function.

[0160] During the negotiation process, the agent exchanges offers and counteroffers with multiple opponent agents until the negotiation meets the termination condition. The negotiation termination condition includes negotiation success or negotiation failure; where negotiation success is defined as: the aggregate utility value of the current counteroffer is greater than the new behavior state, or the expected solution is The aggregate utility value of is greater than the new behavior state, expressed as:

[0161] ;

[0162] in, Desire solution Middle The aggregate utility value of each agent; Indicates a new behavior state; represents the aggregate utility value of the current counter-offer B;

[0163] Negotiation failure is defined as: no new solution can be found , or the negotiation resources are exhausted, such as the behavior state is less than .

[0164] When providing a resource allocation solution, PA evaluates CA's quotation. If the evaluation result is to accept the quotation, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the current CA to allocate resources to it. Otherwise, based on the satisfaction level of PA, a feasible solution set is generated through concession calculation; the expected solution is obtained based on the feasible solution; based on the expected solution, a resource allocation solution, i.e., a counter-offer, is generated until the negotiation termination conditions are met.

[0165] In summary, the algorithm of the present invention includes the following core contents:

[0166] (1) Negotiation information based on fuzzy constraints: Use fuzzy constraints to model negotiation topics to uniformly, flexibly and comprehensively represent negotiation information and solve the problem of information uncertainty and imprecision. By modeling negotiation topics as distributed fuzzy constraint satisfaction problems, constraints between agents are interconnected, and constraints are continuously relaxed through the negotiation process until a solution that satisfies all constraints is generated.

[0167] (2) Construct a unified agent negotiation model: Build a unified multilateral, multi-issue unified negotiation framework to handle the negotiation between consumer users (CA) and edge device providers (PA) to improve the flexibility and efficiency of negotiation and uniformly solve the problem of edge computing resource allocation. At the same time, support the scalability of resources so that the model can maintain efficient expansion and performance in the face of scale changes. The overall negotiation process includes five steps: negotiation topic generation (quotation request), negotiation topic evaluation (solution evaluation), calculation of feasible solutions, quotation generation, and negotiation termination.

[0168] (3) Resource allocation algorithm based on fuzzy rewards: In the edge computing environment where collaboration and competition coexist, a resource allocation algorithm based on fuzzy rewards is proposed. This algorithm needs to balance the interests of each participant and the dynamic needs of the partners in the market environment, while considering long-term returns, so as to achieve results that are more suitable for the dynamic market than fixed strategies and promote the rational allocation of market resources. Among them, the goals of the negotiation agent are expressed using fuzzy functions, which can reflect the preferences of the participants and enhance the rationality of resource allocation. At the same time, the stability index is introduced as one of the reward factors to enhance the algorithm's flexibility in handling various dynamic and random problems.

[0169] In another preferred embodiment, the present invention is secondary developed based on the open source EdgeCloudSim platform. EdgeCloudSim is a popular edge computing simulation environment that provides the facilities, parameters and condition configurations required for the edge computing environment. The platform itself is extended based on CloudSim, so it can be used to simulate the resource environment allocation of edge computing, and simulate a large-scale edge computing infrastructure environment by setting up multiple edge server centers. It has also become an edge computing simulation analysis platform frequently used by scholars.

[0170] This paper studies and expands the existing EdgeCloudSim platform to provide integration of multi-agent development platform, especially to support multi-agent negotiation environment. In the experimental environment, 30 heterogeneous edge servers (PMs) are designed. Each edge server is modeled as having 10 CPU cores, 8GB of RAM, and 1TB of storage space. The CPU performance of 10 PMs is 1000 MIPS, 10 PMs have 2000 MIPS, and the other 10 PMs have 4000 MIPS.

[0171] During the verification and analysis process, the target selection is the aggregation utility value, success rate and other indicators, which are the comparison indicators often used in negotiation. In addition, resource allocation optimization is also the goal of negotiation. Here, the edge server utilization and resource load balancing are analyzed as the goals.

[0172] In terms of algorithm comparison, the negotiation algorithm options are: the Time algorithm that only considers the time factor, the Market algorithm that considers the market factor, the T_M_B algorithm that comprehensively considers the time, market and behavior factors, and the fuzzy negotiation algorithm FN proposed in the present invention.

[0173] like Figure 5As shown, it shows the average aggregate utility value AU obtained by successful negotiation when multiple edge computing centers PA negotiate with multiple users CA in an edge computing environment where the resource demand / supply ratio increases from 0.1 to 1.5. The maximum average aggregate utility value is 2 (i.e., the AU of CA is 1 and the AU of PA is 1). The average AU decreases as the demand supply ratio increases because the edge server has fewer and fewer available resources to meet user-specific requests. It can be seen that the average AU obtained by the FN algorithm of the present invention is the highest compared to other algorithms.

[0174] like Figure 6 As shown, the ratio of successful negotiations is shown when multiple edge computing centers PA negotiate with multiple users CA in an edge computing environment where the resource demand / supply ratio increases from 0.1 to 1.5. When the demand / supply ratio varies in the range of 0.1 to 0.8, resources are relatively sufficient, and the success rate of all negotiation models is greater than 0.9. In addition, it can be found that the T_M_B algorithm and the proposed FN algorithm that consider the opponent's behavior attempt to obtain a satisfactory solution for the opponent. Therefore, these behavior-based models can reduce the timeout failure rate and maintain a high success rate. Among them, the FN algorithm of the present invention achieves the highest success rate compared with other algorithms, because the FN algorithm reduces the failure rate without a solution.

[0175] like Figure 7 As shown, as the resource demand / supply ratio increases, the CPU utilization of the edge server used by the successful request also gradually increases. The FN algorithm of the present invention uses a higher resource utilization of the edge server, which is about 5%-10% higher than other algorithms.

[0176] The FN algorithm of the present invention not only considers the CPU load information of each edge server, but also considers the load balance between edge servers, thus avoiding the problem of processing timeout caused by excessive debt. Figure 8 As shown in Figure 2, the standard deviation of the load changes under different resource demand / supply ratios. Figure 8 It can be seen that the FN algorithm of the present invention is superior to other algorithms, and the standard deviation of the load is much smaller. Compared with other intelligent negotiation algorithms, the FN algorithm of the present invention not only uses the least edge service centers to meet user requests, but also can find appropriate resource utilization to meet the load requirements of the better solution.

[0177] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0178] The resource allocation optimization algorithm of the present invention can adapt to the dynamic changes of resources, improve the efficiency and performance of resource allocation, and meet the service quality requirements of diverse users. Through intelligent negotiation and dynamic resource allocation, while meeting the personalized needs of users, the optimal use of resources is achieved to improve resource utilization and cost-effectiveness, and ensure the stability and reliability of the system.

[0179] Embodiment 2

[0180] like Figure 3 As shown, the second embodiment of the present invention also provides an edge computing resource allocation optimization system based on agent negotiation, including: multiple consumer agents CA and multiple provider agents PA; multiple CAs initiate negotiation requests by broadcasting resource quotes to multiple PAs in the communication area; CAs and PAs continuously exchange quotes and counter-offers until the negotiation is terminated; wherein,

[0181] The consumer agent CA is used to: broadcast its resource allocation negotiation topic request, i.e., quotation, to multiple provider agents PA in the communication area; receive the resource allocation solution provided by PA, i.e., counter-offer; evaluate the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the selected PA to achieve resource allocation; otherwise, combined with the satisfaction level, a new behavior state is obtained through concession calculation; wherein the new behavior state is the satisfaction of the agent with the current resource allocation solution; based on the new behavior state, a feasible solution set is generated; based on the feasible solution set, the optimal solution is selected to obtain the expected solution; based on the expected solution, a new quotation is generated to conduct a new round of negotiation topics until the negotiation termination condition is met;

[0182] The provider intelligent agent PA is used to: receive the quotation from CA for evaluation. If the evaluation result is to accept the quotation, the two parties reach a consensus and the negotiation is successful. The service quality requirement SLA is determined with the current CA and resources are allocated to it. Otherwise, a feasible solution set is generated through concession calculation in combination with the satisfaction level; the expected solution is obtained based on the feasible solution; and a resource allocation solution, i.e., a counter-offer, is generated based on the expected solution until the negotiation termination conditions are met.

[0183] Embodiment 3

[0184] The third embodiment of the present invention also provides an edge computing resource allocation optimization device based on agent negotiation, which includes a memory and a processor, wherein a computer program is stored in the memory, and the computer program can be executed by the processor to implement the edge computing resource allocation optimization method based on agent negotiation as described above.

[0185] Embodiment 4

[0186] The fourth embodiment of the present invention also provides a computer-readable storage medium, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a device where the computer-readable storage medium is located, the edge computing resource allocation optimization method based on agent negotiation as described above is implemented.

[0187] In several embodiments provided in the embodiments of the present invention, it should be understood that the disclosed apparatus and method can also be implemented in other ways. The apparatus and method embodiments described above are merely schematic. For example, the flowcharts in the accompanying drawings show the possible architecture, functions and operations of the apparatus, method and computer program product according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0188] In addition, the functional modules in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0189] If the function is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, an electronic device, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code. It should be noted that in this article, the term "include", "include" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements that are not explicitly listed, or also includes elements inherent to such process, method, article or device. Without more constraints, an element defined by the phrase "comprising a..." does not exclude the existence of other identical elements in the process, method, article or apparatus comprising the element.

[0190] The terms used in the embodiments of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The singular forms "a", "said" and "the" used in the embodiments of the present invention and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.

[0191] It should be understood that the term "and / or" used in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0192] The word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrases "if it is determined" or "if (stated condition or event) is detected" may be interpreted as "when it is determined" or "in response to determining" or "when detecting (stated condition or event)" or "in response to detecting (stated condition or event)", depending on the context.

[0193] The "first\second" mentioned in the embodiments is only to distinguish similar objects, and does not represent a specific order for the objects. It is understandable that the "first\second" can be interchanged with the specific order or sequence where permitted. It should be understood that the objects distinguished by "first\second" can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0194] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An edge computing resource allocation optimization method based on agent negotiation, applied to consumer agent CA, includes the following steps: S1, broadcasts its resource allocation negotiation topic request, i.e., quotation, to multiple provider agents PA in the communication area; S2, receiving the resource allocation solution provided by PA, i.e., counter-offer; S3, evaluate the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the selected PA to achieve resource allocation; otherwise, combined with the satisfaction level, a new behavior state is obtained through concession calculation; wherein the new behavior state is the satisfaction of the intelligent agent with the current resource allocation solution; the satisfaction level is reflected by the constraints of its own satisfaction, the opponent's response information, time limit and market factors; Among them, self-satisfaction is It means that the definition is as follows: ； in, Indicates potential solutions; Indicates The aggregate utility value of each agent; The opponent's response information is through the quotation Counter-offer with the opponent The degree of response between It is defined as: ； in, Indicates the degree of response; Indicates the nth round of bidding; represents the opponent's nth round of counter-offer; n represents the current negotiation round; represents the n-1th round of bidding, i.e. the previous round of bidding; D represents the distance between the bid A and the counter-offer B, which is calculated by the following formula; ； in, Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; represents the distance measure between two fuzzy sets, where the fuzzy set represents the probability distribution of offer A and counter-offer B in each negotiation topic of the agent; The time limit constraint is used to ensure that the negotiation process is carried out within an acceptable time frame and is defined as follows: ； in, Indicates the negotiation time of the current negotiation round; Indicates the negotiation deadline of the current negotiation round; represents a constant, The initial value of ; represents the slope control parameter, when When , the time curve is steeper, and the concession becomes larger as n increases; Market factors evaluate the ratio of market supply and demand by considering the difference between the ratio of the offer and the opponent's concession. If the concession value of the current negotiation round is less than the average concession value, the demand is less than the supply. Market factors is defined as follows: ； in, represents the distance between offer A and counter-offer B in the nth round of negotiation; Indicates the average distance between the offer and the counteroffer in the previous negotiation rounds at the nth negotiation round; S4, generating a feasible solution set according to the new behavior state; S5, selecting the optimal solution according to the feasible solution set and the counter-offer to obtain the expected solution; S6, generating a new quotation based on the expected solution to request a new round of negotiation topics until the negotiation termination condition is met.

2. According to claim 1, a method for optimizing edge computing resource allocation based on agent negotiation is characterized in that When evaluating the resource allocation solution, the aggregate utility value is calculated through the utility function to express the agent's preference for the resource allocation solution; wherein the utility function is expressed as: ； in, Indicates the solution S The aggregate utility value of each agent; Indicates the solution S The fuzzy membership function of the negotiation topics; Indicates the number of negotiation topics being negotiated; Indicates The weight factor of each negotiation topic.

3. According to claim 1, the method for optimizing edge computing resource allocation based on agent negotiation is characterized in that ,The new behavior state is obtained by concession calculation, specifically: Based on the current negotiation cycle's own satisfaction , Response level , Negotiation time and market factors , calculate the agent's concession value , the formula is as follows: ； in, , , , According to their own satisfaction , Response level , Negotiation time and market factors The concession conditions obtained by the constraints; Based on the calculated handicap value , get the new behavior state : ； in, Indicates the current behavior status.

4. According to claim 1, a method for optimizing edge computing resource allocation based on agent negotiation is characterized in that , the formula of the feasible solution set P is as follows: ； in, Indicates a solution; Indicates a new behavior state; Indicates the new intention obtained according to the new behavior state; Indicates A fuzzy constraint network; Indicates the solution S The aggregate utility value of each agent; Indicates the current behavior status.

5. According to claim 1, a method for optimizing edge computing resource allocation based on agent negotiation is characterized in that , the expected solution The formula is: ； in, represents the solution; B represents the counter-offer; P represents the feasible solution set; ) represents the metric function, and its calculation formula is: ； in, Indicates the number of negotiation topics being negotiated; Indicates the solution S The fuzzy membership function of the negotiation topics; represents the weight factor associated with solution satisfaction; Respectively represent the similarity weight factors with solution satisfaction; Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; It represents the distance measure between two fuzzy sets, where the fuzzy sets represent the possibility distribution of offer A and counteroffer B in each negotiation topic.

6. According to claim 1, a method for optimizing edge computing resource allocation based on agent negotiation is characterized in that , the new quotation is: by generating a new quotation element for each object of the current negotiation topic, the new quotation is composed of the new quotation elements, which is expressed as: ； ； in, indicating said new offer; Represents the new quote element of the qth object, that is, the kth agent in the space The marginal likelihood distribution in ; Indicates the total number of objects of the current negotiation topic; is the expected solution; The kth agent of the qth object is in the space Cylinder expansion in ; represents the space of the qth object and the kth agent; Proj represents the space projection function.

7. The method for optimizing edge computing resource allocation based on agent negotiation according to claim 1 is characterized in that ,The negotiation termination conditions include negotiation success or negotiation failure; wherein, Negotiation success is defined as: the aggregate utility value of the current counter-offer is greater than the new behavior state, or the expected solution is The aggregate utility value of is greater than the new behavior state, expressed as: ； in, Desire solution Middle The aggregate utility value of each agent; Indicates a new behavior state; represents the aggregate utility value of the current counter-offer B; Negotiation failure is defined as: no new solution can be found or negotiation resources are exhausted.

8. The method for optimizing edge computing resource allocation based on agent negotiation according to any one of claims 1 to 7, characterized in that ,When PA provides a resource allocation solution, it evaluates the CA’s quotation. If the evaluation result is to accept the quotation, the two parties reach a consensus and the negotiation is successful. The PA determines the service quality requirement SLA with the current CA and allocates resources to it; Otherwise, combined with the satisfaction level of PA, a feasible solution set is generated through concession calculation; The expected solution is obtained based on the feasible solution set; based on the expected solution, a resource allocation solution, i.e., a counter-offer, is generated until the negotiation termination condition is met.

9. An edge computing resource allocation optimization system based on intelligent agent negotiation, characterized in that: include: Multiple consumer agents CA and multiple provider agents PA; Multiple CAs initiate negotiation requests by broadcasting resource quotes to multiple PAs in the communication area; The CA and the PA exchange offers and counteroffers until the negotiation ends; The consumer agent CA is used to: broadcast its resource allocation negotiation topic request, i.e., quotation, to multiple provider agents PA in the communication area; receive the resource allocation solution provided by PA, i.e., counter-offer; evaluate the resource allocation solution; if the evaluation result is to accept the resource allocation solution, the two parties reach a consensus, the negotiation is successful, and the service quality requirement SLA is determined with the selected PA to achieve resource allocation; otherwise, combined with the satisfaction level, a new behavior state is obtained through concession calculation; wherein the new behavior state is the satisfaction of the agent with the current resource allocation solution; based on the new behavior state, a feasible solution set is generated; based on the feasible solution set, the optimal solution is selected to obtain the expected solution; based on the expected solution, a new quotation is generated to conduct a new round of negotiation topics until the negotiation termination condition is met; The provider agent PA is used to: receive the CA's quotation for evaluation. If the evaluation result is to accept the quotation, the two parties reach a consensus and the negotiation is successful. The service quality requirement SLA is determined with the current CA and resources are allocated to it; otherwise, a feasible solution set is generated through concession calculation in combination with the satisfaction level; the expected solution is obtained based on the feasible solution set; a resource allocation solution, i.e., a counter-offer, is generated based on the expected solution until the negotiation termination condition is met; the satisfaction level is reflected by the constraints of the party's own satisfaction, the opponent's response information, time limit and market factors; Among them, self-satisfaction is It means that the definition is as follows: ； in, Indicates potential solutions; Indicates The aggregate utility value of each agent; The opponent's response information is through the quotation Counter-offer with the opponent The degree of response between It is defined as: ； in, Indicates the degree of response; Indicates the nth round of bidding; represents the opponent's nth round of counter-offer; n represents the current negotiation round; represents the n-1th round of bidding, i.e. the previous round of bidding; D represents the distance between the bid A and the counter-offer B, which is calculated by the following formula; ； in, Indicates The fuzzy constraints on the offer A in the negotiation topic; Indicates The fuzzy constraint on the counter-offer B in the negotiation topic; represents the distance measure between two fuzzy sets, where the fuzzy set represents the probability distribution of offer A and counter-offer B in each negotiation topic of the agent; The time limit constraint is used to ensure that the negotiation process is carried out within an acceptable time frame and is defined as follows: ； in, Indicates the negotiation time of the current negotiation round; Indicates the negotiation deadline of the current negotiation round; represents a constant, The initial value of ; represents the slope control parameter, when When , the time curve is steeper, and the concession becomes larger as n increases; Market factors evaluate the ratio of market supply and demand by considering the difference between the ratio of the offer and the opponent's concession. If the concession value of the current negotiation round is less than the average concession value, the demand is less than the supply. Market factors is defined as follows: ； in, represents the distance between offer A and counter-offer B in the nth round of negotiation; Indicates the average distance between the offer and the counteroffer in the previous negotiation rounds at the nth negotiation round.

Images