AUTONOMOUS DISTRIBUTED CONTROL INVOLVING CONSTRAINT ON RESOURCES

• US 20080082981A1
• Filed: 10/02/2007
• Published: 04/03/2008
• Est. Priority Date: 10/02/2006
• Status: Active Grant

• Current Assignee: Japan Aerospace Exploration Agency
• Original Assignee: Japan Aerospace Expioraion Agency Institute of Space and Astronautical Science
• Inventors: Tsuda, Yuichi, Nakaya, Koji, Mori, Osamu, Kawaguchi, Junichiro
• Granted Patent: US 8,347,298 B2
• Time in Patent Office: Days
• US Class Current: 718/104
• CPC Class Codes: H04L 47/70 Admission control; Resource allocation

Abstract

Disclosed is a distributed control method of cooperatively allowing a control strategy for satisfying a constraint on an aggregated resource and simultaneously achieving or maintaining a performance of a system to be determined in a real-time manner by any one of a plurality of entities which have the highest need for the resource at any given time, without providing a specified management entity. The proposed novel method or “card game scheme” of the present invention is designed to dynamically carry out a function of satisfying a constraint on an aggregate resource and simultaneously achieving or maintaining a performance of the entire system, based on a distributed control with high flexibility, wherein an “agent”, i.e., a software having a function of performing autonomous determination and control, is set in each of the entities, to allow the entities to dynamically take turns to perform a processing of determining a strategy for resource allocation, without specifying the agent serving as a management entity having an authority to determine the strategy, so as to provide a processing method with high versatility.

Claims

1. A method for use in a system including:

a plurality of entities (agents) each adapted to exhibit a specific performance white consuming a resource;

a mechanism adapted to supply the resource to each of said entities;

a mechanism provided in each of said entities to accept the resource; and

means provided in each of said entities to carry out a committing function of effecting said specific performance, or an actuation device provided in each of said entities to effect said specific performance, said method being designed to satisfy a constraint on an aggregated resource to be in input from said resource supply mechanism into the entire system, and simultaneously control the respective specific performances of said entities, said method comprising the steps of;

providing a network adapted to connect between respective ones of said entitles so as to allow data about the resource consumption and the specific performance in each of said entities to be exchanged therebetween, and communication means having a function of transmitting anti receiving said data;

assigning to each of said entities a target index with which said specific performance is to be brought into conformity within an allowable range around said target index;

assigning to each of said entities an interval reference index during which the resource is input into said entity;

providing means for carrying out a function of measuring the specific performance or the resource consumption in each of said entitles;

providing to each of said entitles means for carrying out a software function of autonomously and cooperatively determining a strategy; and

allowing said entities to take turns to perform said software function.

2. The method as defined in claim 1, wherein said software function includes, when each of said entities is assigned with an initial index of said interval reference index, and dynamically shifted between two roles of a server and a client;

dynamically performing four actions of said server consisting of bidding, evaluating, committing and shifting, or an action of said client consisting of reporting, so as to allow said entities to distributedly perform a processing of determining a strategy for the entire system.

3. The method as defined in claim 2, wherein the software function in any selected one of said entities whenever said role of said server is delegated thereto according to said shifting action, includes:

committing the resource consumption in said selected entity to bring said specific performance into conformity to said target index or maintain said specific performance at said target index;

performing said bidding action to collect first data about a difference index between an actual specific performance and said target index in each of the remaining entitles and second data about a difference index between said assigned interval reference index allowing the resource to be consumed at a predetermined index, and an actual number of time-frames in an elapsed time period after the latest turn-on instance in each of said remaining entities;

performing said evaluating action to select one of said remaining entities which is to be allowed to commit the resource consumption in a subsequent time-frame; and

performing said shifting action to delegate an authority to determine said strategy.

4. The method as defined in claim 2, wherein the software function in each of the remaining entities other than said server includes, in response to a tendering request from said server, reporting a status of the specific performance and the resource input in said entity.

5. The method as defined in claim 2, wherein the software function in each of the remaining entities other than said server includes, in response to a tendering request from said server, reporting a status of the specific performance and the resource input in said entity.

6. The method as defined in claim 2, wherein the software function in each of the remaining entities other than said server includes, in response to a tendering request from said server, reporting first data about a difference index between an actual specific performance and said target index in each of the remaining entitles, and second data about a difference index between said assigned reference interval index allowing the resource to be consumed at a predetermined index, and an actual number of time-frames in an elapsed time period after the latest turn-on instance in each of said remaining entities, in the form of three types of Cards consisting of:

an On Card which is issued when said actual specific performance is below a lower limit of said allowable range;

a Coast Card which is issued when said actual specific performance falls within said allowable range, said Coast Card being indicative of a difference index between said assigned interval reference index and an actual time-frame number in an elapsed time period after a latest turn-on; and

a Skip Card which is issued when said actual specific performance is above an upper limit index of said allowable range, said Skip Card being indicative of no need for positively consuming the resource.

7. The method as defined in claim 6, wherein said software function in said server includes, during said evaluating action based on said collected first anti second data which contains one or more On Cards, selecting, as a next server, one of the remaining entity which has issued the On Card having a largest difference index, and, during said evaluating action based on said collected first and second data which contains no On Cards and one or more Coast Card, selecting;

as a next server, one of the remaining entity which has issued the Coast Card having the largest difference index.

8. The method as defined in claim 6, wherein said software function in each of the remaining entities other than said server includes avoiding issuing the Skip Card on a case-by-case basis.

9. The method as defined in claim 6, wherein said software function in, said server includes, during said committing action, clearing an index of said Coast Card in each of said remaining entities to zero, and subsequently continuing said committing action with reference to said initially assigned interval reference index.

10. The method as defined in claim 6, wherein said software function in said server includes, during said committing action, clearing an index of said Coast Card in each of said remaining entities to zero, and then adaptively controlling each of said remaining entities based on an actual performance thereof after an operation of adjusting and updating said interval reference index with reference to said pre-cleared index of said Coast Card.

11. The method as defined in claim 1, wherein said software function in said server includes, during said committing action, integrating respective interval reference indexes updated according to said adaptive control in said remaining entitles, and multiplying each of said interval reference indexes in said remaining entitles by a certain coefficient which allows said updated interval reference indexes through said multiplication to be matched to an aggregate resource specified based on said integrated index.

12. The method as defined in claim 1, which includes the steps of normalizing respective resource consumption indexes to be determined for said entities, at a same index.

13. The method as defined in claim 1, wherein:

said system is a heating system and/or a cooling system;

said resource is electric power; and

said specific performance in each of said entities is temperature.

14. The method as defined in claim 1, wherein:

said system is a data input-output system;

said resource is a segmented data traffic; and

said specific performance in each of said entities is a utilization efficiency of a buffer device.

15. The method as defined in claim 11, wherein said network of said system is installed concurrently with and along a line for supplying said resource.