Expert system capable of meeting real time constraints

• US 5,412,753 A
• Filed: 11/27/1992
• Issued: 05/02/1995
• Est. Priority Date: 11/29/1991
• Status: Expired due to Fees

• Current Assignee: Alcatel NV (Nokia Corporation)
• Original Assignee: Alcatel NV (Nokia Corporation)
• Inventors: Alston, Andrew, Delory, Jean-Michel, Laasri, Hassan, Gandara, Marc
• Primary Examiner(s): Downs, Robert W.
• Application Number:
  US07/982,311
• Time in Patent Office: 886 Days
• Field of Search: 395/12, 395/916, 395/917, 395/575, 395/650, 395/11, 395/51, 395/54, 395/77, 371/16.1, 371/18
• US Class Current: 706/11
• CPC Class Codes: G06F 11/2257 using expert systems
• CPC Class Codes: G06N 5/043 Distributed expert systems; Blackboards
• CPC Class Codes: H04L 41/0631 using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis
• CPC Class Codes: H04L 41/0663 Performing the actions predefined by failover planning, e.g. switching to standby network elements
• CPC Class Codes: H04L 41/16 using machine learning or artificial intelligence
• CPC Class Codes: H04L 43/0817 by checking functioning
• CPC Class Codes: Y10S 706/916 Electronic or computer, internal or network, circuit
• CPC Class Codes: Y10S 706/917 Communication

Abstract

An expert system capable of meeting real time constraints for solving a problem on the basis of a flow of input messages concerning the operation and the status of a network of machines, this status changing with time, includes a filter module detecting in real time significant events on the basis of input messages, each significant event being a manifestation of a problem, as judged by a human expert. A database stores all input messages, the significant events defined by the filter module, intermediate data produced by the expert system, data constituting a model of the network of machines and archives relating to problems processed previously, whether solved or not. A reasoning module comprises a set of rules divided into a plurality of subsets called specialists and each comprising knowledge provided by a human expert required to process part of a problem, specialists being capable of executing in parallel without blocking each other, and a command system for selecting and executing specialists according to heuristic rules provided by a human expert and making a decision or declaring that a decision is impossible at the end of a predetermined time period in respect of a problem represented by significant events communicated to the reasoning module by the filter module. A graphical interface is provided for human/expert system communications. An expert assistant module supplies a human expert with data stored in the database.

Claims

1. Expert system for solving problems on the basis of a flow of input messages concerning the operation and the status of a network of machines while meeting real time constraints, said status changing with time, said system comprising:

filter means for detecting in realtime significant events on the basis of input messages, each significant event being a manifestation of a problem, as judged by a human expert;

a database for storing;

all input messages,

the significant events detected by the filter means,

intermediate data produced by the expert system,

a model of the network of machines, and

archives relating to problems processed previously, whether solved or not;

reasoning means for reasoning responsive to said filter means and to said database, said reasoning means comprising;

correlating means for determining whether a significant event detected in realtime by said filter means can be correlated with significant events previously detected by said filter means to thereby establish a new or updated common problem associated with a plurality of interrelated significant events having a possible common cause;

a plurality of specialists for processing part of the common problem, each comprising a subset of rules and knowledge provided by a human expert and being capable of executing in parallel without blocking each other; and

command means responsive to each significant event detected in realtime by said filter means and also responsive to each new or updated common problem associated with a plurality of interrelated significant events established by said correlating means for selecting and executing specialists according to heuristic rules provided by a human expert and making a decision or declaring that a decision is impossible at the end of a predetermined time period in respect of a problem represented by significant events communicated to said reasoning means by said filter means;

a graphical interface for human/expert system communications; and

expert assistant means for supplying a human expert with data stored in said database.

2. Expert system according to claim 1 wherein said filter means comprises means for:

distinguishing between different types of input messages;

counting messages of predetermined different types during sliding time periods of fixed duration; and

detecting a significant event each time a count of messages exceeds a fixed threshold.

3. Expert system according to claim 1 wherein:

said input messages, said expert assistant means, said filter means, said command means, said specialists, said heuristic rules and said network model are implemented by a computer operating under the control of software written in an object-oriented language; and

said database is an object-oriented database.

4. Expert system according to claim 1 wherein at least some items of data stored in said database are associated with a respective problem stamp;

and said correlating means comprises means for using said problem stamp as an access key to access a respective said item of data.