Abstract
A class of sequential Monte Carlo estimation is frequently called particle filter. This filter belongs to the Bayesian strategy for estimation, where a non-linear and non-Gaussian assumptions can be applied. Here, the Tsallis’ distribution, from the non-extensive thermo-statistics, is used to design the best likelihood operator. Therefore, no previous likelihood operator is assumed. The new filter formulation will be named as non-extensive particle filter (NEx-PF). The distribution estimated by the NEx-PF can compute the standard form of the central limit theorem, as well as the Levy-Gnedenko central limit theorem. The q-calculus formalism is employed to generalize some definitions and properties.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Here, the word distribution will have the same meaning as probability density function (PDF).
References
Campos Velho, H. F., Furtado, H. C. M.: Adaptive particle filter for stable distribution. In: Integral Methods in Science and Engineering, C. Constanda and P. J. Harris (eds.), Birkhäuser, New York, NY (2011), pp. 47–57.
Ernst, T. 2003. A Method for q-Calculus. Journal of Nonlinear Mathematical Physics, 10,
Hernández Torres, R., Scarabello, M., Campos Velho, H. F., Chiwiacowsky, L. D., Soterroni, A., Gouvea, E., Ramos, F. M.: A Hybrid Method using q-Gradient to Identify Structural Damage. In: XXXVI Iberian Latin American Congress on Computational Methods in Engineering (CILAMCE), 2015, Rio de Janeiro (RJ), Brazil (2015), pp. 1–14.
Jackson, F. H.: On q-functions and a certain difference operator. Trans. Roy Soc. Edin., 46, 253–281 (1908).
Jackson, F. H.: On q-definite integrals. Quart. J. Pure and Appl. Math., 41, 193–203 (1910).
Jackson, F. H. 1910b. q-difference equations. American Journal of Mathematics, 32, 307–314 (1910).
Kac, V., Cheung, P.: Quantum Calculus. Springer-Verlag, New York (2002).
Kaipio, J., Somersalo, E.: Statistical and Computational Inverse Problems, In Applied Mathematical Sciences 160 – series, Springer-Verlag, (2004).
Nolan, J. P.: Stable Distributions: Models for Heavy Tailed Data, Boston, MA: Birkhäuser (2005).
Ristic, B., Arulampalam, S., Gordon, N.: Beyond the Kalman Filter, Boston, Artech House (2004).
Schön, T. B.: Estimation of Nonlinear Dynamic Systems: Theory and Applications, Dissertations no. 998 – Linköping Studies in Science and Technology (2006).
Soterroni, A. C., Galski, R. L., Ramos, F. M.: The q-gradient vector for unconstrained continuous optimization problems. In Operations Research Proceedings, B. Hu, K. Morasch, S. Pickl and M. Siegle (eds.), Springer, Berlin, Heidelberg (2011), pp. 365–370.
Tsallis, C.: Possible Generalization of Boltzmann-Gibbs Statistics, J. Statistical Phys., 52, 479–487 (1988).
Tsallis, C.: Nonextensive statistics: theoretical, experimental and computational evidences and connections, Braz. J. Phys., 29, 1–35 (1999).
C. Tsallis, S. M. D. Queiros, Nonextensive statistical mechanics and central limit theorems I - Convolution of independent random variables and q-product, in Complexity, Metastability and Non-extensivity (CTNEXT 07) (Editors: S. Abe, H. Herrmann, P. Quarati, A. Rapisarda, and C. Tsallis), American Institute of Physics, 2007.
Umarov, S., Tsallis, C., Steinberg, S.: On a q-Central Limit Theorem Consistent with Nonextensive Statistical Mechanics. Milan J. Math., 76, 307–328 (2008).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix: Non-extensive Tsallis’ Thermostatistics
Appendix: Non-extensive Tsallis’ Thermostatistics
A non-extensive form to the entropy can be expressed as [Ts88]:
and the q-expectation of an observable is given by
Some properties are derived to the non-extensive entropy S q.
-
1.
If q → 1:
$$\displaystyle \begin{aligned} S_1=k \sum_{i=1}^{N_p}p_i \ln p_i~, \end{aligned} $$(3.28)$$\displaystyle \begin{aligned} O_1=\sum_{i=1}^{N_p}p_i O_i~. \end{aligned} $$(3.29) -
2.
The non-extensive entropy has a positive value: S q ≥ 0.
-
3.
Non-extensivity:
$$\displaystyle \begin{aligned} S_q(A+B)=S_q(A)+S_q(B)+(1-q)S_q(A)S_q(B) \end{aligned} $$(3.30)$$\displaystyle \begin{aligned} O_q(A+B) = O_q(A)+O_q(B) + (1-q) [ O_q(A)S_q(B)+O_qS_q(A) ]~. \end{aligned} $$(3.31) -
4.
Maximum for the S q under the constraint \(O_q = \sum _i p_i^q \epsilon _i\) (canonical ensemble):
$$\displaystyle \begin{aligned} p_i=\frac{1}{Z_q}[1-\beta(1-q)\epsilon_i]^{1/(1-q)} \end{aligned} $$(3.32)where 𝜖 i is the energy for the state-i, O q = U q is the non-extensive form to the internal energy and the normalization factor Z q (partition function), for 1 < q < 3, is given by
$$\displaystyle \begin{aligned} Z_q=\left[ \frac {\pi} {\beta(1-q)} \right]^{1/2} \frac{\varGamma [(3-q)/2(q-1)]}{\varGamma[1/(q-1)]}~. \end{aligned} $$(3.33)For q = 1, the normalization is written as
$$\displaystyle \begin{aligned} p_i=e^{\beta \epsilon_i}/Z_1~. \end{aligned} $$(3.34)
Rights and permissions
Copyright information
© 2019 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Araújo, A.S., Furtado, H.C.M., de Campos Velho, H.F. (2019). q-Calculus Formalism for Non-extensive Particle Filter. In: Constanda, C., Harris, P. (eds) Integral Methods in Science and Engineering. Birkhäuser, Cham. https://doi.org/10.1007/978-3-030-16077-7_3
Download citation
DOI: https://doi.org/10.1007/978-3-030-16077-7_3
Published:
Publisher Name: Birkhäuser, Cham
Print ISBN: 978-3-030-16076-0
Online ISBN: 978-3-030-16077-7
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)