Estimation of Distribution algorithm for sensor selection problems

Author

Naeem, M. ; Lee, D.C.

Author_Institution

Sch. of Eng. Sci., Simon Fraser Univ., Burnaby, BC, Canada

fYear

2010

fDate

10-14 Jan. 2010

Firstpage

388

Lastpage

391

Abstract

In this paper, we apply Estimation-of-Distribution Algorithms (EDAs) to the problem of selecting a set of k sensors from m sensors for the purpose of parameter estimation. Unlike other evolutionary algorithms, in EDAs a new population of individuals in each iteration is generated without crossover and mutation operators; instead, a new population is generated based on a probability distribution, which is estimated form the best selected individuals of previous iteration. Our results indicate that EDA is a good candidate for solving the sensor selection problems.

Keywords

evolutionary computation; parameter estimation; probability; estimation of distribution algorithm; evolutionary algorithms; parameter estimation; probability distribution; sensor selection problems; Chemical sensors; Electronic design automation and methodology; Estimation error; Evolutionary computation; Genetic mutations; Maximum likelihood estimation; Parameter estimation; Probability distribution; Sensor systems; Virtual manufacturing; EDA; Sensor Selection;

fLanguage

English

Publisher

ieee

Conference_Titel

Radio and Wireless Symposium (RWS), 2010 IEEE

Conference_Location

New Orleans, LA

Print_ISBN

978-1-4244-4725-1

Electronic_ISBN

978-1-4244-4726-8

Type

conf

DOI

10.1109/RWS.2010.5434261

Filename

5434261