Combining Population-Based Search with Nonlinear Optimization to Assist Decision Making in Presence of Multiple Objectives

PDF

Authors
  1. Pinter, J.D.
  2. Gammon, M.A.
Corporate Authors
Defence R&D Canada - Atlantic, Dartmouth NS (CAN);Pinter Consulting Services
Abstract
Canada Command and DRDC CORA sponsored a study regarding the solution of non-linear optimization problems with multiple objectives. Many decision problems are expressed by non-linear functions, including practical cases which require a simulation approach. The current study begins with a review of the problem and some related literature. It then describes one methodology based on a Population-Based Optimisation (PBO) for the initial screening of the decision space. The Lipschitz Global Optimizer (LGO) program is examined for use as the non-linear, global optimization solver. Finally, the multiple objective aspects are handled using a weighted average or aggregate approach. This combination may possibly be useful for suitable military or other topical decision problems of practical interest. The current study can be considered as a first examination of the problem, and further research into other methodologies and approaches, as well as testing with suitable and tractable problems should be considered.

Il y a un résumé en français ici.

Report Number
DRDC-ATLANTIC-TM-2010-192 — Technical Memorandum
Date of publication
01 Mar 2012
Number of Pages
62
DSTKIM No
CA040896
CANDIS No
801976
Format(s):
Electronic Document(PDF)

Permanent link

Document 1 of 1

Date modified: