Monte carlo filter using the genetic algorithm operators
- 1 August 1997
- journal article
- research article
- Published by Taylor & Francis in Journal of Statistical Computation and Simulation
- Vol. 59 (1) , 1-23
- https://doi.org/10.1080/00949659708811843
Abstract
We consider the generalized state space model (GSSM) which is an extension of the state space model to the non-Gaussian and non-linear model. There are serious problems in the GSSM approach because of the need for numerical integration over a state space. A Monte Carlo method for filtering and smoothing, called the Monte Carlo Filter (MCF), has been proposed to overcome this numerical problem. It has been pointed out that there exists a close relationship between the MCF and the genetic algorithm (GA) and that an essential structure involved in the MCF is quite similar to that in the GA. In this study, we try to replace the step of the prediction by the mutation and crossover operators in the GA, and demonstrate their performance as the system noise. We furthermore propose a smoothing algorithm in which a massively simple parallel procedure plays an important role. The proposed method is first applied to a simple problem and then to a seasonal adjustment for quarterly data sets in order to illustrate its broad applicability.Keywords
This publication has 12 references indexed in Scilit:
- Monte Carlo Filter and Smoother for Non-Gaussian Nonlinear State Space ModelsJournal of Computational and Graphical Statistics, 1996
- Bayesian Inference in Cyclical Component Dynamic Linear ModelsJournal of the American Statistical Association, 1995
- Applied state space modelling of non-Gaussian time series using integration-based Kalman filteringStatistics and Computing, 1994
- Posterior Mode Estimation by Extended Kalman Filtering for Multivariate Dynamic Generalized Linear ModelsJournal of the American Statistical Association, 1992
- A Monte Carlo Approach to Nonnormal and Nonlinear State-Space ModelingJournal of the American Statistical Association, 1992
- Nonlinear multiparameter optimization using genetic algorithms: Inversion of plane‐wave seismogramsGeophysics, 1991
- Non-Gaussian seasonal adjustmentComputers & Mathematics with Applications, 1989
- Bayesian statistical inference to remove periodic noise in the optical observations aboard a spacecraftApplied Optics, 1988
- Dynamic Generalized Linear Models and Bayesian ForecastingJournal of the American Statistical Association, 1985
- A New Approach to Linear Filtering and Prediction ProblemsJournal of Basic Engineering, 1960