Multiobjective Functions Optimization For Task Allocation Mapping To Multicomputer Nodes
Abstract
The Naval Air Warfare Center - Aircraft Division Warminster, Software and Computer Technology Division, is developing software to allocate tasks to multicomputer nodes such that multiple objectives are optimized. The eventual goal of this project is to allow the decision maker to have design optimization capabilities available to her/him when designing complex system. Two key aspects of the allocation problem are discussed: (1) The techniques used to treat multiple objectives simultaneously and (2) The techniques used to perform the search for the optimal set of model parameters. The report discusses current research reported in the literature and makes several suggestions for continuing the current effort. The current technique being used for treating multiple objectives simultaneously is to use a weighted sum of each of the individual objective functions as the objective function for the search. It is suggested that a vector of objective functions approach be implemented and experimentally compared to the weighted sum approach. Two algorithms are being used in the software: the genetic algorithm and the simulated annealing algorithm.... Complex system engineering, Genetic algorithm, Multicomputer nodes, Multiple objectives, Neural network, Optimization, Searching simulated annealing algorithm, Vector of objective functions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 07, 1992
- Accession Number
- ADA267071
Entities
People
- Jesse Williams
Organizations
- Naval Air Warfare Center Aircraft Divison