Computer-aided decision making with a symbolic spreadsheet

• US 5,182,793 A
• Filed: 06/09/1992
• Issued: 01/26/1993
• Est. Priority Date: 06/30/1989
• Status: Expired due to Term

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Irrgang, Michael E., Kirchner, John A., Alexander, Rhonda L.
• Primary Examiner(s): MacDonald, Allen R.
• Application Number:
  US07/896,731
• Time in Patent Office: 231 Days
• Field of Search: 395/13, 395/10, 364/419, 364/51, 364/54
• US Class Current: 706/13
• CPC Class Codes: G06F 40/18 of spreadsheets form-filling G06F40/174
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

An apparatus and method for assisting persons in making decisions, using a computer programmed with artificial intelligence techniques. Real world objects and events pertaining to a particular domain are represented in a knowledge base. Best choices for solving problems are made according to the application of rules, which may be applied absolutely, comparatively, by weight, or ordered, according to methods selected by the user. The invention also permits the user to select from among various decision making strategies and permits the user to observe the effects of choices in hypothetical scenarios.

Claims

1. In a computer system having a stored system of rules and a knowledge base of facts, events, and programming, an improved method for using said computer to display a problem scenario and modify the scenario according to a best choice among alternative choices in a particular solution domain, comprising the steps of:

representing a real world problem to be solved as at least one data object;

generating a problem scenario from a number of data objects, wherein said objects are arranged as a symbolic spreadsheet in which cells of said spreadsheet are frames representing said objects and their attributes and wherein said cells are linked by relationship attributes of said objects;

displaying said problem scenario to a user;

generating a pool of candidate solutions to said problem scenario;

determining a candidate selection strategy for evaluating said candidates, wherein said strategy defines a method for applying rules;

evaluating said candidates, using said candidate selection strategy to access and apply a combination of hard elimination rules, soft elimination rules, and comparison rules from a stored system of rules, to each of said candidate solutions until a best candidate is selected;

assigning a solution value representing said best candidate to variables of said problem scenario;

modifying said scenario to substitute said solution value, such that a new scenario reflecting the effects of said solution value is generated, wherein all objects having attributes affected by said solution value are modified; and

displaying said modified scenario to a user.

2. The method of claim 1, wherein said candidate generating step further comprises updating the characteristics of said candidates in accordance with predetermined data or user input.

3. The method of claim 1, wherein said strategy is selected in accordance with user input.

4. The method of claim 1, wherein said evaluating step comprises evaluating said candidates using a hierarchial method.

5. The method of claim 1, wherein said evaluating step comprises evaluating said candidates using a rule weighting method.

6. The method of claim 5, wherein said rule weighting method includes a step for short circuiting said evaluation step by determining that a current best candidate cannot be surpassed.

8. The method of claim 1, wherein said candidate evaluating step further comprises using soft elimination rules.

9. The method of claim 1, wherein said candidate evaluating step further comprises using comparison rules.

10. The method of claim 1, wherein said candidate evaluating step further comprises using scoring rules.

11. The method of claim 1, wherein said candidate evaluating step further comprises using stop rules.

12. The method of claim 1, and further comprising the step of using said best choice in a final action.

13. The method of claim 1, wherein said candidate generating step further comprises updating the characteristics of said alternative choices in accordance with predetermined data or user input.

14. The method of claim 1, wherein said computer provides more than one method of applying said rules during said step of evaluating.

7. The method of claim 29, wherein said rule weighting method learns best candidate selection by adjusting said rules.