Patent · US Expired

Multi-objective optimization method

US7139741B1 · kind B1 · utility

10Cited by
2References
22Claims
0Family size

Assignee

Inventor

Key dates

Filing dateJul 31, 2003
Grant dateNov 21, 2006
Priority date
Expiry dateOct 15, 2024

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06N5/04
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

A mathematical programming method in which a common decision space is defined by identifying decision variables, and a set of objective functions is developed as piecewise defined functions. Each objective function is dependent on at least one of the decision variables and corresponds to a goal. The objective function has a plurality of pieces, such that the objective function has only one piece for each combination of decision variables. A value function is developed in additive preference form based on weights and the objective functions. The value function is optimized by searching through the decision variable values to find the values of the decision variables that maximize the developed value function. An action dictated by the decision variable values can then be executed.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.