Available online at www.sciencedirect.

com

Procedia IUTAM 5 (2012) 199 – 203

IUTAM Symposium on 50 Years of Chaos: Applied and Theoretical

Chaos and its applications

Kazuyuki Aiharaa*
a
Institute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan

Abstract

The concept of deterministic chaos has been making large impacts both on science and on engineering. Possible applications of
chaotic dynamics are considered from a viewpoint of chaos engineering.

©2012
© 2011Published
Publishedbyby Elsevier
Elsevier Ltd.Ltd. Peer-review
Selection under responsibility
and/or Peer-review of TakashiofHikihara
under responsibility and Tsutomu
Takashi Hikihara Kambe Kambe
and Tsutomu
Open access under CC BY-NC-ND license.
Keywords: chaos; chaos engineering; chaotic neuron; chaotic neural network

1. Introduction

The concept of deterministic chaos has been greatly influencing not only science but also engineering, technology,
and even arts along with substantial progress in our understanding of deterministic chaos since 1970's. In 1990, we
established ''chaos engineering'' at Japan Electronic Industry Development Association where chaos engineering is
defined as generic studies of theoretical and technological foundations for possible applications of deterministic chaos,
fractals, and complex systems[1].
In this article, I will review research on applications of chaotic dynamics in chaos engineering [1, 2], and preview a
future direction[3].

2. Chaos Engineering

In Japan, many practical applications of deterministic chaos, even for household appliances, have been developed in
various fields of engineering and technology. Actually, studies of nonlinear dynamics in engineering disciplines have
been steadily progressing in Japan over a half century. In particular, Yoshisuke Ueda in the Department of Electrical
Engineering at Kyoto University found a strange attractor, which he called ''the broken-egg chaotic attractor,'' in the
following Duffing-van der Pol mixed type equation in 1961[4,5]:
d 2x dx
dt 2

 P 1  Jx 2
dt

 x3 B cos Zt. (1)

Further, a strange attractor in the following Duffing equation was named ''the Japanese attractor'' [6-8]:

*
Corresponding author.
E-mail address: aihara@sat.t.u-tokyo.ac.jp.

1877-7058 © 2012 Published by Elsevier Ltd. Selection and/or Peer-review under responsibility of Takashi Hikihara and Tsutomu Kambe
Open access under CC BY-NC-ND license. doi:10.1016/j.piutam.2012.06.027
200 Kazuyuki Aihara / Procedia IUTAM 5 (2012) 199 – 203

d 2x dx
2
k  x3 B cos Zt. (2)
dt dt
Electrical and electronic circuits are typical fields in which rich chaotic phenomena have been reported. One of the
reasons is because a variety of reliable nonlinear elements are readily available in electrical and electronic circuits.
In 1927, van der Pol and van der Mark heard noise corresponding to deterministic chaos in an electrical system
composed of a neon glow lamp, a resistor, a variable condenser, and D.C. and A.C. power sources [9]. Since then,
chaos has been detected both experimentally and numerically in many electrical and electronic circuits like the series
RLC circuit with a varactor diode [10], Chua's circuit, and Shinriki's circuit [11, 12]. It should be noted that both the
Duffing equation and the Duffing-van der Pol mixed type equation are also models of electrical circuits with an
nonlinear inductance and an nonlinear resistance, respectively.
These studies have confirmed that chaos can be actually and easily generated in nonlinear electrical and electronic
circuits. In particular, periodic forcing with an A.C. power source is common in circuits, which provides a variety of
examples for driven chaos in non-autonomous systems. The ubiquity of chaos in circuits conversely implies that
chaotic dynamics of mathematical models may be implemented in real-world systems by these circuits beyond
computer simulations toward possible applications.
Influences of deterministic chaos on engineering are not limited to electrical and electronic engineering. For
example, R. E. Kalman noted random behavior in nonlinear sampled-data control systems [13]. Further, characteristics
of deterministic chaos are greatly affecting basic concepts of engineering such as prediction, control, computation,
information, and optimization [1-3, 14, 15]. Therefore, ''chaos engineering'' has been actively researched and
developed in Japan for more than twenty years [1, 2, 16-20].
Applications of deterministic chaos are conceptually classified as applicable chaos and applied chaos [1, 16-18].
Here, whereas applied chaos means direct applications of known properties on deterministic chaos to concrete
examples in real-world systems, applicable chaos aims at generalization and systematization of mathematical
structures of chaotic dynamics common to such concrete complex phenomena toward possible further applications of
chaos in the future [1, 17].
Among many chaos-engineering applications so far developed, one of very promising areas is nonlinear time series
analysis [20-24] on complex time series data from real-world systems. Examples of such time series analysis include
predictive models derived directly from observed data such as time series of temperature in the iron-making blast
furnace [25], daily peak loads in electric power grids [26], Japanese vowels [27-29], and irregular and complicated roll
motion of a flooded ship in waves [30-32].
Chaos engineering has produced even consumer products with chaotic dynamics for household appliances. Chaotic
dynamics with 1/f-like power spectra was utilized for temperature control in kerosene fan heaters and for direction
control of air streams in air conditioners to provide useful fluctuations that are expected to be comfortable to humans
[1, 19, 20]. Chaos engineering has also been successfully applied to dishwashers [1, 19, 33, 34]. A usual dishwasher is
equipped with a one-link nozzle. By changing the conventional dishwasher nozzle design from a one-link structure to a
two-link structure where the center of the gravity in the smaller second link is slightly different from the center of its
rotation, movement of the nozzle changes from periodic to chaotic rather than quasiperiodic [34]. This is, as an
analogy, similar to what happens in conservative pendulum systems where a single pendulum behaves periodically but
a double pendulum does chaotically. Chaotic loci of water jet holes in the two-link dishwasher cover homogeneously a
wider area and achieve better washing performance than do periodic ones in the conventional one-link dishwasher.

3. Chaotic Neural Networks

The nonlinear dynamics in biological neurons generates and propagates action potentials. Such nonlinear dynamics
of neurons has been studied experimentally with squid giant axons and theoretically with the Hodgkin-Huxley
equations [35]. In fact, chaos and various bifurcations to chaos in neurons were found both with the Hodgkin-Huxley
equations and with squid giant axons [36-43].
Moreover, a chaotic neuron model that qualitatively reproduces chaotic dynamics observed in squid giant axons can
be described by a simple 1-dimensional map [44-46] as follows:
x(t  1) F ( x(t )) { kx(t )  Df {x(t )}  a, (3)
 Kazuyuki Aihara / Procedia IUTAM 5 (2012) 199 – 203 201

where x(t) is the internal state at discrete time t, k is the decay factor with |k|<1, a is the bias including both the
threshold and the temporally constant input, and output y(t) is obtained by y(t)=f{x(t)} with, for example, the logistic
function f{x(t)}=1/{1+exp(-x(t)/İ)}.
The chaotic neuron model of eq.(3) was extended to neural network models called chaotic neural networks [44, 45,
47]. The chaotic neural networks composed of the chaotic neurons produce various kinds of spatio-temporal dynamics
with abilities of parallel distributed processing [2, 44, 45, 47]. The dynamics of the ith chaotic neuron in the model of
chaotic neural networks is described as follows:
M
[i (t  1) ¦ v A (t )  k [ (t ),
j 1
ij j e i (4)

N
Ki (t  1) ¦ w y (t )  k K (t ),
j 1
ij j f i (5)

9 i (t  1) Dgi ^yi (t )`  kr9 i (t )  Ti . (6)

The output yi(t+1) of the ith neuron is calculated by transforming the total internal state ȟi(t+1)+Și(t+1)+ȗi(t+1) to the
output value through the output function fi as follows:
yi (t  1) f i ^[ i(t  1) K i (t  1) 9 i(t  1)`. (7)
Here ȟi(t), Și(t), and ȗi(t) are the internal states for the externally applied spatio-temporal inputs from neurons
outside the network, the spatio-temporal feedback inputs from the constituent neurons within the network, and the
refractoriness, respectively, Aj(t) is the strength of the input from the jth external neuron at discrete time t, vij and wij
are synaptic weights from the jth external neuron and from the jth constituent neuron to the ith constituent neuron,
respectively, ke, kf, and kr are the decay factors for the external inputs, the feedback inputs, and the refractoriness,
respectively with |ke|, |kf|, |kr| ӊ 1, gi(yi) is the refractory function that represents the strength of the refractory effects
caused by the output yi, and Į and și are the refractory scaling parameter and the resting threshold value, respectively.
The ''analysis by synthesis'' or a ''constructive approach'' is a general methodology for complex systems, a typical
example of which is the brain. This constructive approach to the brain is composed of the following processes. First, a
certain function of biological neurons and neural networks is mathematically modeled on the basis of experimental
results. The chaotic dynamics of squid giant axons and its modeling with the chaotic neuron model is such an example.
Next, a mathematical model of a higher brain function is analysed by mathematical models like the chaotic neural
networks to elucidate their possible mechanisms. If such a model can mathematically reproduce a target function of
the brain, the mathematical model is technologically implemented by software or hardware for possible applications.
Dynamical systems with chaotic dynamics enables complex computations fundamentally with real numbers [2, 48].
To exploit such a computational ability of deterministic chaos, chaotic elements should be implemented with devices
that can faithfully realize real numbers. Analog electronic circuits can actually handle real numbers at least in principle
because the state variables such as voltages, currents, and time are continuous, although noise is inevitably
superimposed [48].
An important characteristics of deterministic chaos is that simple nonlinear systems can frequently generate very
complicated behavior and possibly useful functions. The advantage of simple nonlinear properties necessary for a
genesis of deterministic chaos is that they can be implemented by simple analog hardware. Neurons may effectively
utilize this kind of analog implementation of deterministic chaos through the evolution.
Moreover, artificial implementation of the CNN model with analog electronic circuits is plausible because the
equations of the CNN model are simple enough and can be easily and directly implemented with various existing
analog electronic circuits[48-50] by learning from the pioneering works on chaos in electrical and electronic circuits
explained in the Section 2.

4. Conclusion

Analog circuits with chaotic dynamics such as the chaotic neural networks can be interpreted as contemporary-aged
revival of analog computers in former days [16]. A hybrid computing system composed of present digital computers
and such chaotic analog circuits can be a candidate of a future computing system [3, 48-50].
202 Kazuyuki Aihara / Procedia IUTAM 5 (2012) 199 – 203

Acknowledgements

This research is supported by the Aihara Innovative Mathematical Modelling Project, the Japan Society for the
Promotion of Science (JSPS) through the "Funding Program for World-Leading Innovative R&D on Science and
Technology (FIRST Program)," initiated by the Council for Science and Technology Policy (CSTP).

References

[1] Aihara K, Katayama R. Chaos engineering in Japan. Communications of the ACM 1995;38(11):103–7.
[2] Aihara K. Chaos engineering and its application to parallel distributed processing with chaotic neural networks. Proceedings of the IEEE
2002;90(5):919–30.
[3] Aihara K, editor. Theme Issue on Theory of Hybrid Dynamical Systems and its Applications to Biological and Medical Systems.
Philosophical Transactions of the Royal Society A 2010;368(1930).
[4] Horiuchi K, Koga T, editors. Special Issue on Engineering Chaos. Trans. Institute of Electronic, Information and Communication
Engineerings 1990;E73(6):757-863.
[5] Abraham R, Ueda Y, editors, The chaos avant-garde -memories of the early days of chaos theory. Singapore: World Scientic; 2000.
[6] Ruelle D. Strange attractors. Math. Intelligencer 1980;2:126-37.
[7] Endo T, Saito T. Chaos in electrical and electronic circuits and systems. The Transactions of the IEICE 1990;E73(6):763-71.
[8] Ueda Y. The road to chaos. Aerial Press; 1992.
[9] van der Pol B, van der Mark J. Frequency demultiplication. Nature 1927;120:363-4.
[10] Testa J, Pe'rez J, Jeries C. Evidence for universal chaotic behavior of a driven nonlinear oscillator. Phys. Rev. Lett 1982;48:714-7.
[11] Chua LO, We CW, Hang A, Zhong G-Q. A universal circuit for studying and generating chaos -Part I: Routes to chaos. IEEE Trans. on
Circuits and Systems-I: Fundamental Theory and Applications 1993;40(10):732-44.
[12] van Wyk MA, Steeb W-H. Chaos in electronics. Dordrecht : Kluwer Academic Publishers; 1997.
[13] Kalman RE. Nonlinear aspects of sampled-data control systems. In: Fox J, editor. Proc. Symp. on Nonlinear Circuit Analysis VI, New
York: Polytechnic Institute of Brooklyn; 1956, p. 273–313.
[14] Yamaguti M, editor. Towards the harnessing of chaos. Amsterdam: Elsevier, Science Publishers B.V.; 1994.
[15] Grebogi C, Yorke JA, editors. The impact of chaos on science and society. Tokyo: United Nations University Press; 1997.
[16] Aihara K, editor. Neural, fuzzy and chaotic computing -introduction to a new generation of analog computing. Tokyo: Ohm-sha; 1993 (in
Japanese).
[17] Aihara K, editor. Applied chaos and applicable chaos. Tokyo: Science-sha; 1994 (in Japanese).
[18] Aihara K, Katayama R. A special issue on chaos engineering. J. Intelligent & Fuzzy Systems 1997; 5(1): 1-2.
[19] Aihara K, Iokibe T, editors. Chaos application systems. Tokyo: Asakura; 1995 (in Japanese).
[20] Aihara K. Introduction to chaos. Tokyo: Air University Press; 2001 (in Japanese).
[21] Parker TS, Chua LO. Practical numeriacal algorithms for chaotic systems. Berlin: Springer; 1989.
[22] Ott E, Sauer T, Yorke JA, editors. Coping with chaos: analysis of chaotic data and the exploration of chaotic systems. New York: John
Wiley & Sons, Inc.; 1994.
[23] Abarbanel HDI. Analsis of observed chaotic data. Berlin: Springer; 1995.
[24] Kanz H, Schreiber T. Nonlinear time series analysis. Cambridge: Cambridge University Press; 1997.
[25] Miyano T, Kimoto S, Shibuta H, Nakashima K, Ikenaga Y, Aihara K. Time series analysis and prediction on complex dynamical behavior
observed in a blast furnace. Physica D 2000;135:305-30.
[26] Mizukami Y, Nishimori T, Okamoto J, Aihara K. Forecasting daily peak load by a deterministic prediction meshod with the Gram-
Schmidt orthonormalization. Trans. Inst. Electrical Eng. Jpn. 1995;115-C:792-7 (in Japanese).
[27] Tokuda I, Tokunaga R, Aihara K. A simple geometrical structure underlying speech signals of the Japanese vowel /a/. Int. J. Bifurc and
Chaos 1996;6(1):149-60.
[28] Miyano T, Nagami A, Tokuda I, Aihara K. Detecting nonlinear determinism in voiced sounds of Japanese vowel /a/. Int. J. Bifurc. and
Chaos 2000;10(8):1973-9.
[29] Tokuda I, Miyano T, Aihara K. Surrogate analysis for detecting nonlinear dynamics in normal vowels. Journal of the Acoustical Society
of America 2001;110(6): 3207-17.
[30] Murashige S, Aihara K. Experimental study on chaotic motion of a flooded ship in waves. Proc. R. Soc. Lond. A 1998;454(1978):2537-53.
[31] Murashige S, Yamada T, Aihara K. Nonlinear analyses of roll motion of a flooded ship in waves. Philosophical Transactions of the Royal
Society, Series A 2000;358:1793-812.
[32] Mitsubori K, Aihara K. Delayed-feedback control of chaotic roll motion of a flooded ship in waves. Proceedings of the Royal Society of
London, Series A 2002;458:2801-13.
 Kazuyuki Aihara / Procedia IUTAM 5 (2012) 199 – 203 203

[33] Nomura H, Naito E, Wakami N, Kondo S, Aihara K. Analyses of chaotic behavior of a 2-link nozzle in a dishwasher. Proc. of Japan-
U.S.A. Symposium on Flexible Automation 1994:231-4.
[34] Nomura H, Wakami N, Aihara K. Time series analyses on behavior of a 2-link nozzle in a dishwasher. Trans. IEICE 1995;J78-A(6):678-
85 (in Japanese).
[35] Hodgkin AL, Huxley AF. A quantitative description of membrane current and its application to conduction and excitation in nerve. J.
Physiol. (London) 1952;117:500-544.
[36] Aihara K, Matsumoto G, Ikegaya Y. Periodic and non-periodic responses of a periodically forced Hodgkin-Huxley oscillator. J. Theor.
Biol. 1984; 109: 249-69.
[37] Matsumoto G, Aihara K, Ichikawa M, Tasaki A. Periodic and nonperiodic responses of membrane potentials in squid giant axons during
sinusoidal current stimulation. J. Theor. Neurobiol. 1984;3:1-14.
[38] Aihara K, Matsumoto G, Ichikawa M. An alternating periodic-chaotic sequence observed in neural oscillators. Phys. Lett. A 1985; 111(5):
251-5.
[39] Aihara K, Matsumoto G. Chaotic oscillations and bifurcations in squid giant axons. In: Holden AV, editor. Chaos, Manchester :
Manchester University Press, Princeton: Princeton University Press; 1986, p. 257-69.
[40] Aihara K, Numajiri T, Matsumoto G, Kotani M. Structures of attractors in periodically forced neural oscillators. Phys. Lett. A 1986;116:
313-7.
[41] Matsumoto G, Aihara K, Hanyu Y, Takahashi N, Yoshizawa S, and Nagumo J. Chaos and phase locking in normal squid axons. Phys.
Lett. A 1987;123:162-6.
[42] Mees A, Aihara K, Adachi M, Judd K, Ikeguchi T, Matsumoto G. Deterministic prediction and chaos in squid axon response. Phys. Lett.
A:1992; 169, 41-45.
[43] Aihara K. Chaos in axons. In: Arbib MA, editor. The handbook of brain theory and neural networks. 1st ed. Cambridge, Massachusetts:
The MIT Press; 1995, p.183-5.
[44] Aihara K. 1989 Chaotic neural networks. In: Bifurcation Phenomena in nonlinear systems and theory of dynamical systems, RIMS
Kokyuroku 710, Research Institute for Mathematical Sciences, Kyoto University; 1989, p.145-63.
[45] Aihara K, Takabe T, Toyoda M. Chaotic neural networks. Phys. Lett. A 1990;144(6/7): 333-40.
[46] Aihara K, Suzuki H. Theory of hybrid dynamical systems and its applications to biological and medical systems. Phil. Trans. Roy. Soc. A
2010;368(1930); 4893-914.
[47] Adachi M, Aihara K. Associative dynamics in a chaotic neural network. Neural Networks 1997;10(1):83-98.
[48] Horio Y, Aihara K. Analog computation through high-dimensional physical chaotic neuro-dynamics. Physica D 2008;237(9):1215-25.
[49] Horio Y, Aihara K, Yamamoto O. Neuron-synapse IC chip -Set for large-scale chaotic neural networks. IEEE Trans. Neural Networks
2003; 14(5):1393-404.
[50] Horio Y, Ikeguchi T, Aihara K. A mixedn analog/digital chaotic neuro-computer system for quadratic assignment problems. Neural
Networks 2005;18(5-6):505-513.