GB/Z 185.5-2026Artificial intelligence - Agent interconnection - Part 5: Agent discovery (English PDF)

Foreword

This document is a standard or guiding technical document. This document conforms to GB/T 1.1-2020 "Standardization Work Guidelines Part

1.Structure and Drafting Rules of Standardization Documents". Drafting is scheduled. This document is Part 5 of GB /Z 185 "Artificial Intelligence - Interconnection of Intelligent Agents". GB /Z 185 has already published the following parts. --Part

5 Agent Discovery

1.Scope This document establishes the process for agent discovery for agent interconnection. This document applies to the design, implementation, and testing of interconnected and collaborative solutions for artificial intelligence agents.

4.Abbreviations The following abbreviations apply to this document. API. Application Programming Interface GUI. Graphical User Interface LUI. Language User Interface

6 Agent Discovery Methods

6.1 Discovery Process Based on Agent Discovery Service The agent description manager should provide an agent discovery service and an agent discovery interface to respond to agent discovery requests. The process diagram is shown in Figure 1.

Note. In some systems, the agent discovery service and the agent description and management system are implemented as separate software systems. The discovery service should be integrated with the agent... The description management synchronizes the intelligent agent description data to ensure the accuracy and timeliness of the intelligent agent description information. The discovery process based on the agent discovery service is as follows:

a) The agent sends an agent discovery request to the agent discovery service (step 1 in Figure 1). 1) Discovery conditions can be generated by the agent or set manually; 2) The discovery criteria should preferably include the natural language descriptions defined in Table 1 of GB /Z 185.4-2026 as necessary conditions, and combinations thereof are also acceptable. Includes one or more other information such as the agent's name and identification code; 3) The agent discovery service should provide at least one query interface, such as API, GUI or LUI.

b) Upon receiving the request, the agent discovery service searches for agents that meet the requirements among the published agents (step 2 in Figure 1). 1) The result set may contain 0, 1 or more entries, with 1 entry corresponding to 1 agent; 2) Any entry in the result set should contain at least the name of the agent or any other information that can be used for compliance checks; Example. The intelligent agent discovery service analyzes query conditions (such as natural language descriptions of functions, performance, status, security constraints, or payment information) to identify queries. The intent is inquired and then matched with the target intelligent agent to determine the target intelligent agent that meets the conditions. 3) Based on business needs, the agent discovery service should only return configurations that allow discovery, according to the discoverable configurations of registered agents. Discovered agents do not return agents configured as undiscoverable; 4) Depending on business needs, the agent discovery service should adhere to the availability requirements of registered agents, such as paid use or availability only for specific purposes. Information such as the usage of specific user groups.

c) The agent discovery service returns a result set to the discovery request (step 3 in Figure 1).

d) The compliance of the agents listed in the agent check result set with the current business requirements (step 4 in Figure 1). 1) If the returned agent meets the business needs, it is advisable to initiate interaction with the target agent as needed; 2) If no agent description is returned or the returned agent does not meet business requirements, the query conditions can be modified and the query can be resubmitted. Request for information [see

e) The agent interacts based on the business matching results (step 5 in Figure 1).

6.2 Discovery Process Based on Pre-defined Information The agent discovery process based on pre-set information is shown in Figure 2. Figure

2 Discovery process based on pre-set information The discovery process based on pre-defined information is as follows:

a) The agent uses the query criteria to search for agents in the local preset information (step 1 in Figure 2), and returns the result set. 1) Query criteria can be generated by an intelligent agent or set manually; 2) The local lookup process can be implemented by the agent provider or planned and executed by the agent.

b) Pre-configured agent information should meet business security requirements and may come from any one of the following sources. 1) Pre-configured by the intelligent agent provider; 2) When an agent accesses the cached result set returned by the agent discovery service, the agent should check or store the cached result set according to business requirements. The timeliness of the certificate description (such as periodically confirming the availability of cached agents with the agent discovery service); 3) Agent user configuration; 4).well-known address.

Note..well-known is a standardized directory on a web server used to store specific types of files (such as website introductions, configurations, certificates, etc.). Books, etc.

c) When the query result set is not empty, the agent checks the conformity of the agent in the result set to the current business (Figure 2) according to