Failure diagnosis system, image forming apparatus, computer readable medium and data signal

• US 7,676,703 B2
• Filed: 01/16/2007
• Issued: 03/09/2010
• Est. Priority Date: 06/09/2006
• Status: Expired due to Fees

• Current Assignee: Fuji Xerox Company Limited (Xerox Holdings Corp.)
• Original Assignee: Fuji Xerox Company Limited (Xerox Holdings Corp.)
• Inventors: Satonaga, Tetsuichi, Uwatoko, Koki, Yasukawa, Kaoru
• Primary Examiner(s): Baderman; Scott T
• Assistant Examiner(s): PATEL, KAMINI B
• Application Number:
  US11/653,400
• Publication Number:
  US 20080010522A1
• Time in Patent Office: 1,148 Days
• Field of Search: 714/48, 714/26
• US Class Current: 714/48
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

A failure diagnosis system includes a first database, a second database, an observation-information acquiring unit, a determination-probability calculating unit and an occurrence-probability calculating unit. A first cause-and-effect network stored in the first database stochastically represents a relationship between failure-type candidates and first observation information, which have cause-and-effect relationships with the failure-type candidates. Each second cause-and-effect network stored in the second database stochastically represents relationships between second observation information and failure-cause candidates. The observation-information acquiring unit acquires the first and second observation information from a diagnosed system. The determination-probability calculating unit calculates a determination probability of each failure-type candidate based on the first observation information and the first cause-and-effect network. The occurrence-probability calculating unit calculates occurrence probabilities of the failure-cause candidates for each failure-type candidate based on the second observation information and the second cause-and-effect networks.

Claims

1. A failure diagnosis system comprising:

a first database that stores a first cause-and-effect network, wherein the first cause-and-effect network stochastically represents a relationship between a plurality of failure-type candidates and first observation information, which have cause-and-effect relationships with the respective failure-type candidates, and

a second database that stores a plurality of second cause-and-effect networks for the respective failure-type candidates, wherein each of the second cause-and-effect networks stochastically represents relationships between second observation information and the failure-cause candidates;

an observation-information acquiring unit that acquires the first observation information and the second observation information from a diagnosed system;

a determination-probability calculating unit that calculates a determination probability of each failure-type candidate based on the first observation information acquired by the observation-information acquiring unit and the first cause-and-effect network;

an occurrence-probability calculating unit that calculates occurrence probabilities of the failure-cause candidates for each failure-type candidate based on the second observation information acquired by the observation-information acquiring unit and the second cause-and-effect networks;

a failure-cause notifying unit that notifies at least one of the failure-cause candidates as a failure cause, based on the determination probabilities of the respective failure-type candidates calculated by the determination-probability calculating unit and the occurrence probabilities of the respective failure-cause candidates calculated by the occurrence-probability calculating unit; and

a multiplication unit that multiplies (i) the determination probability of each failure-type candidate by (ii) the respective occurrence probabilities of the failure-cause candidates for the failure-type candidate, wherein,

the failure-cause notifying unit notifies the at least one of the failure-cause candidates based on multiplication results calculated by the multiplication unit.

2. The system according to claim 1, further comprising:

a sum-total calculating unit that calculates a sum total of the multiplication results, calculated by the multiplication unit, of each failure-cause candidate common to the second cause-and-effect networks, wherein;

the failure-cause notifying unit notifies at least part of the failure-cause candidates in descending order of the total sums calculated by the sum-total calculating unit.

3. The system according to claim 1, further comprising:

a multiplication unit that calculates P(Mi)×

P(Cij) (i=1, 2, . . . n; and

j=1, 2, . . . m), where;

n denotes natural number;

m denotes natural number;

Mi denotes the respective failure-type candidates;

Cij denotes the respective failure-cause candidates for the corresponding failure-type candidate Mi;

P(Mi) denotes the respective determination probabilities of the failure-type candidates Mi; and

P(Cij) denotes the respective occurrence probabilities of the failure-cause candidates Cij, wherein;

the failure-cause notifying unit notifies the at least one of the failure-cause candidates based on P(Mi)×

P(Cij) calculated by the multiplication unit.

4. The system according to claim 3, further comprising:

a sum-total calculating unit that calculates

$\underset{f}{\sum <br><br>}<br><br><br><br>P<br><br>\left(\mathrm{Mi}\right)\times <br><br>P<br><br>\left(\mathrm{Cij}\right),$

 

wherein;

the failure-cause notifying unit notifies at least part of the failure-cause candidates in descending order of

$\underset{f}{\sum <br><br>}<br><br><br><br>P<br><br>\left(\mathrm{Mi}\right)\times <br><br>P<br><br>\left(\mathrm{Cij}\right)$

 

calculated by the sum-total calculating unit.

5. An image forming apparatus comprising:

a print engine unit that forms an image on a sheet of paper; and

the system according to claim 1.

6. A computer readable storage medium storing a program causing a computer to execute a process for a failure diagnosis, the process comprising:

acquiring first observation information and second observation information from a diagnosed system;

calculating a determination probability of each failure-type candidate based on the acquired first observation information and a first cause-and-effect network, wherein the first cause-and-effect network stochastically represents a relationship between the plurality of failure-type candidates and the first observation information, which have cause-and-effect relationships with the respective failure-type candidates;

calculating occurrence probabilities of failure-cause candidates for each failure-type candidate based on the acquired second observation information and a plurality of second cause-and-effect networks for the respective failure-type candidates, wherein each second cause-and-effect network stochastically represents relationships between the second observation information of the failure-type candidate and the failure-cause candidates;

notifying at least one of the failure-cause candidates as a failure cause, based on the calculated determination probabilities of the respective failure-type candidates and the calculated occurrence probabilities of the respective failure-cause candidates; and

multiplying (i) the determination probability of each failure-type candidate by (ii) the respective occurrence probabilities of the failure-cause candidates for the failure-type candidate, wherein, the at least one of the failure-cause candidates is notified based on the multiplication results of multiplying the determination probability of each failure-type candidate by the respective occurrence probabilities of the failure-cause candidates for the failure-type candidate.