Case-based inference processing method

• US 5,243,689 A
• Filed: 06/11/1991
• Issued: 09/07/1993
• Est. Priority Date: 06/15/1990
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Yoshiura, Hiroshi, Mori, Fumihiko
• Primary Examiner(s): Fleming, Michael R.
• Assistant Examiner(s): Downs, Robert W.
• Application Number:
  US07/713,761
• Time in Patent Office: 819 Days
• Field of Search: 395/51, 395/76, 395/919
• US Class Current: 706/12
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: Y10S 706/919 Designing, planning, programming, CAD, CASE

Abstract

In a computer system having a storage unit for storing at least one of a program and knowledge used for solving a problem, an input/output unit for inputting a problem to be presently solved, and an execution unit for obtaining a proposed solution of one of the problem to be presently solved and a partial problem of the problem, by using at least one of the program and knowledge in the storage unit, an inference processing method includes the steps of storing a modification case generated upon modification of a solution, calculating a portion, unable to be solved by the proposed solution obtained by the execution unit, of one of the problem to be solved and the partial problem, searching for a modification case containing a problem similar to the unsolved portion, from the modification cases generated upon modification of solutions to problems, correlating the searched modification case with the proposed solution obtained by the execution unit, and modifying the proposed solution obtained by the execution unit in accordance with the correlation and using the modified result as one of an overall solution and an intermediate solution of the problem to be solved presently.

Claims

1. An inference processing method for a computer system having a storage unit for storing at least one of a program and knowledge used for solving a problem, an input/output unit for inputting a problem to be presently solved, and an execution unit for obtaining a proposed solution of the problem to be presently solved and a partial problem of the problem, by using at least one of the program and knowledge in the storage unit, said inference processing method comprising the steps of:

storing at least one modification case generated upon modifications of prior solutions, into said storage unit, said modification case including a solution before modification, a modification process for obtaining the solution, and a partial problem solved by the modification process;

inputting said problem to be solved, from said input/output unit;

obtaining by said execution unit a proposed solution of one of said inputted problem to be solved and a partial problem of said inputted problem, by using said program and knowledge in said storage unit;

identifying a portion of said problem to be solved and said partial problem, unable to be solved by said proposed solution obtained by said execution unit;

searching for a modification case containing a partial problem similar to said unsolved portion, from at least one modification case generated upon modification of the prior solutions, and stored in said storage unit;

correlating said searched for modification case with said proposed solution obtained by said execution unit; and

automatically modifying said proposed solution obtained by said execution unit in accordance with said correlation and using said modified result as one of an overall solution and an intermediate solution of said problem to be solved, and wherein, said modifying step includes applying said modification process in said modification case to one of said proposed solution obtained by said execution unit and a solution in said solution case searched by a solution case retrieving unit, in accordance with said correlation, and using the applied result as one of the overall solution and the intermediate solution of said problem to be solved.

2. An inference processing method according to claim 1, wherein said automatically modifying step further includes modifying one of said proposed solution obtained by said execution unit and said solution in solution case searched by said solution case retrieving unit so as to match said solution after the modification in said modification case, in accordance with said correlation, and using the modified result as one of the overall solution and the intermediate solution of said problem to be solved.

3. An inference processing method according to claim 2, further comprising the steps of:

displaying by said input/output unit one of said proposed solution obtained by said execution unit and said solution in said solution case searched by said solution case retrieving unit; and

when the displayed solution is modified by a user, storing a modification case including a modifying portion of the solution before modification, a modified portion of the solution after the modification, and a partial problem solved by the modification.

4. An inference processing method according to claim 1, further comprising the steps of:

displaying by said input/output unit one of said proposed solution obtained by said execution unit and said solution in said solution case searched by said solution case retrieving unit; and

when the displayed solution is externally modified, storing a modification case including a modifying portion of the solution before modification, a modification process, and a partial problem solved by the modification process.

5. An inference processing method according to claim 1, wherein in searching for said modification case, a similarity between two partial problems is judged based on common subject of two sets of characteristic features which represent each of the partial problems.

6. An inference processing method according to claim 1, further comprising a step of modifying again one of said overall solution and said intermediate solution obtained in accordance with said modification case so as to apply said modified result to one of said problem to be solved presently and said partial problem.

7. An inference processing method according to claim 1, further comprising a step of when searching for said modification case results in a plurality of modification cases, evaluating solutions obtained from said plurality of modification cases and using an optimum modification case.

8. An inference processing method according to claim 7, wherein said modification case includes priority order information, when the optimum modification case is selected by evaluating said solutions, the priority order of said optimum modification case is made high and a modification case is selected in accordance with said priority order.

9. An inference processing method according to claim 7, wherein when the optimum modification case is selected by evaluating said solutions, a priority order of said optimum modification case is made high and a modification case having a higher priority order is preferentially used.

10. An inference processing method for an inference system having a solution case data base for storing a solution case including a problem and at least one of a proposed solution of the problem and a way to solve the problem, an input/output unit for inputting a problem to be solved, a solution case retrieving unit for searching for a solution case having a problem similar to the problem to be solved and a partial problem of the problem, from the solution case data base, and a solution case modifying unit for, respectively, modifying one of the proposed solution and the way to solve the searched for solution case, to obtain one of an overall solution and an intermediate solution of the problem to be solved, said inference processing method comprising the steps of:

storing at least one modification case generated upon modification of prior solutions in a storage unit, said modification case including a solution before modification, a modification process for obtaining the solution, and a partial problem solved by the modification process;

inputting a problem to be solved, from said input/output unit;

searching by said solution case retrieving unit a solution case having a problem similar to one of the inputted problem to be solved and a partial problem of the inputted problem, from said solution case data base;

when said solution case modifying unit modifies the solution of said searched solution case, searching for a modification case having a partial problem similar to a portion corresponding to a difference between one of the problem to be solved and the partial problem, and the problem contained in said searched solution case, from said at least one modification case in said storage unit, and correlating said searched modification case with said searched solution case; and

automatically modifying said proposed solution obtained by said searched solution case in accordance with said correlation and using said modified result as one of an overall solution and an intermediate solution of said problem to be solved, and wherein said modifying step includes applying said modification process in said modification case to one of said proposed solution obtained by an execution unit and a solution in said solution case searched by said solution case retrieving unit, in accordance with said correlation, and using the applied result as one of the overall solution and the intermediate solution of said problem to be solved.

11. An inference processing method according to claim 10, wherein said automatically modifying step further includes modifying one of said proposed solution obtained by said execution unit and said solution in said solution case searched by said solution case retrieving unit so as to match said solution after the modification in said modification case, in accordance with said correlation, and using the modified result as one of the overall solution and the intermediate solution of said problem to be solved presently.

12. An inference processing method according to claim 11, further comprising the steps of:

displaying by said input/output unit one of said proposed solution obtained by said execution unit and said solution in said solution case searched by said solution case retrieving; and

when the displayed solution is modified by a user, storing a modification case including a modifying portion of the solution before modification, a modified portion of the solution after the modification, and a partial problem solved by the modification.

13. An inference processing method according to claim 10, further comprising the steps of:

displaying by said input/output unit one of said proposed solution obtained by said execution unit and said solution in said solution case searched by said solution case retrieving unit; and

when the displayed solution is externally modified, storing a modification case including a modifying portion of the solution before modification, a modification process, and a partial problem solved by the modification process.

14. An inference processing method according to claim 10, wherein in searching for said modification case, a similarity between two partial problems is judged based on common subject of two sets of characteristic features which represent each of the partial problems.

15. An inference processing method according to claim 10, further comprising the step of modifying again one of said overall solution and said intermediate solution obtained in accordance with said modification case so as to apply said modified result to one of said problem to be solved presently and said partial problem.

16. An inference processing method according to claim 10, further comprising a step of, when searching for said modification case results in a plurality of modification cases, evaluating solutions obtained from said plurality of modification cases and using an optimum modification case.

17. An inference processing method according to claim 16, wherein said modification case includes priority order information, when the optimum modification case is selected by evaluating said solutions, the priority order of said optimum modification case is made high and a modification case is selected in accordance with said priority order.

18. An inference processing method according to claim 16, wherein when the optimum modification case is selected by evaluating said solutions, a priority order of said optimum modification case is made high and a modification case having a higher priority order is preferentially used.