Browse Prior Art Database

(Textual) Objective Modeling for Implicit elicitation of decision maker?s multi-objective preferences.

IP.com Disclosure Number: IPCOM000242160D
Publication Date: 2015-Jun-21
Document File: 5 page(s) / 270K

Publishing Venue

The IP.com Prior Art Database

Abstract

Decision makers are often asked to choose among a variety of alternatives. In such problems there are several objectives that need to be optimized in parallel, and a solution that optimizes all of them does not exist. Instead there is a set of Pareto-optimal alternatives. Each alternative is a compromise with a trade-off over the values of the various objectives, and the decision maker needs to decide which alternative is subjectively best for her. However, a process aimed at assisting the decision maker to properly evaluate and rank the alternatives requires extra information from the decision maker in order to extract any sort of utility function. We offer a method that automatically extracts information to support a decision maker to find her optimal policy. The output of the method is used to provide baseline to the decision maker on the importance of the objectives, rank the alternatives and provide comparable framework with other decision makers. The method utilizes textual information provided by the decision maker (either professional, social or other) and textual information on the objectives (formal definitions, research papers, Wikipedia articles and other sources of information). The method is not limited to individual human decision makers, but is also applicable to extract organizational significance of the objectives out of the corpus of the organization published documents.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 34% of the total text.

Page 01 of 5

( (Textual

Textual) ) preferences .

Decision makers are often asked to choose among a variety of alternatives. These alternatives are typically different solutions to a multi objective optimization problem (Single objective problems in real world situations are extremely rare). In such problems there are several objectives that need to be optimized in parallel, and a solution that optimizes all of them does not exist. For example, a minister for environmental issues would like to adapt a policy that on one hand reduces the cost of food, increasing bio diversity and reducing CO2 emission or a consumer that would like to select a hotel which is of maximal review score, minimal nightly rate and as close as possible to the city center. An ideal alternative that optimize all the goals does not exists. Instead there is a set of Pareto-optimal alternatives. Each alternative is a compromise with a trade-off over the values of the various objectives, and the decision maker needs to decide which alternative is subjectively best for her. Note, that an alternative which is not in the Pareto front Defineā€¦ is ignored because if all objective are properly modeled, there will always be another alternative (within the Pareto-optimal set) which is superior to that alternative. However, a process aimed at assisting the decision maker to properly evaluate and rank the alternatives requires extra information from the decision maker in order to extract any sort of utility function.

    We offer a method that automatically extracts information to support a decision maker to find her optimal policy. The output of the method is used to provide baseline to the decision maker on the importance of the objectives, rank the alternatives and provide comparable framework with other decision makers. The method utilizes textual information provided from the decision maker corpus(either professional, social or other) and textual information on the objectives (formal definitions, research papers, Wikipedia articles and other sources of information) in order to provide the functionality. The method is not limited to individual human decision makers, but is also applicable to extract organizational significance of the objectives out of the corpus of the organization published documents.

The core idea of the invention is to utilize free text documents to represent both each objective and the decision makers. From these texts, the system automatically extracts the utility function that corresponds to the importance of each objective to the decision makers.

This output can be used as follows:

2 Enable a complete ranking of the various alternatives based on the subjective utility extracted by the method.

2 Feedback to the decision maker on the extracted objectives' significance to allow fine tuning and calibration.

2 Comparison and clustering with other decision makers based on their subjective utility.


2 Compound of integrated utility based on many individual inputs.

Obje...