-
Universal power laws in the threshold network model: A theoretical analysis based on extreme value theory
Authors:
A. Fujihara,
M. Uchida,
H. Miwa
Abstract:
We theoretically and numerically investigated the threshold network model with a generic weight function where there were a large number of nodes and a high threshold. Our analysis was based on extreme value theory, which gave us a theoretical understanding of the distribution of independent and identically distributed random variables within a sufficiently high range. Specifically, the distribu…
▽ More
We theoretically and numerically investigated the threshold network model with a generic weight function where there were a large number of nodes and a high threshold. Our analysis was based on extreme value theory, which gave us a theoretical understanding of the distribution of independent and identically distributed random variables within a sufficiently high range. Specifically, the distribution could be generally expressed by a generalized Pareto distribution, which enabled us to formulate the generic weight distribution function. By using the theorem, we obtained the exact expressions of degree distribution and clustering coefficient which behaved as universal power laws within certain ranges of degrees. We also compared the theoretical predictions with numerical results and found that they were extremely consistent.
△ Less
Submitted 19 November, 2009;
originally announced November 2009.
-
On the principle of minimum growth rate in multiplicatively interacting stochastic processes
Authors:
Akihiro Fujihara,
Toshiya Ohtsuki,
Hiroshi Yamamoto
Abstract:
A method of moment inequalities is used to derive the principle of minimum growth rate in multiplicatively interacting stochastic processes(MISPs). When a value of a power-law exponent at the tail of probability distribution function exists in a range $0 < s \le 1$, a first-order moment diverges and an equality for a growth rate of systems breaks down. From the estimate of inequalities, we newly…
▽ More
A method of moment inequalities is used to derive the principle of minimum growth rate in multiplicatively interacting stochastic processes(MISPs). When a value of a power-law exponent at the tail of probability distribution function exists in a range $0 < s \le 1$, a first-order moment diverges and an equality for a growth rate of systems breaks down. From the estimate of inequalities, we newly find a conditional inequality which determines the growth rate, and then the exponent in $0 < s \le 1$.
△ Less
Submitted 8 August, 2006;
originally announced August 2006.
-
Log-normal distribution in growing systems with weighted multiplicative interactions
Authors:
Akihiro Fujihara,
Satoshi Tanimoto,
Toshiya Ohtsuki,
Hiroshi Yamamoto
Abstract:
Many-body stochastic processes with weighted multiplicative interactions are investigated analytically and numerically. An interaction rate between particles with quantities $x, y$ is controlled by a homogeneous symmetric kernel $K(x, y) \propto x^{w} y^{w}$ with a weight parameter $w$. When $w<0$, a method of moment inequalities is used to derive log-normal type tails in probability distributio…
▽ More
Many-body stochastic processes with weighted multiplicative interactions are investigated analytically and numerically. An interaction rate between particles with quantities $x, y$ is controlled by a homogeneous symmetric kernel $K(x, y) \propto x^{w} y^{w}$ with a weight parameter $w$. When $w<0$, a method of moment inequalities is used to derive log-normal type tails in probability distribution functions. The variance of log-normal distributions is expressed in terms of the weight $w$ and interaction parameters. When interactions are weak and a growth rate of systems is small, in particular, the variance is in proportion to the growth rate. This behavior is totally different from that of one-body stochastic processes, where the variance is independent of the growth rate. At $w>0$, Monte Carlo simulations show that the processes end up with a winner-take-all state.
△ Less
Submitted 10 May, 2006; v1 submitted 25 November, 2005;
originally announced November 2005.
-
Asymptotic analysis of the model for distribution of high-tax payers
Authors:
Hiroshi Yamamoto,
Toshiya Ohtsuki,
Akihiro Fujihara,
Satoshi Tanimoto,
Keizo Yamamoto,
Sasuke Miyazima
Abstract:
The z-transform technique is used to investigate the model for distribution of high-tax payers, which is proposed by two of the authors (K. Y and S. M) and others. Our analysis shows an asymptotic power-law of this model with the exponent -5/2 when a total ``mass'' has a certain critical value. Below the critical value, the system exhibits an ordinary critical behavior, and scaling relations hol…
▽ More
The z-transform technique is used to investigate the model for distribution of high-tax payers, which is proposed by two of the authors (K. Y and S. M) and others. Our analysis shows an asymptotic power-law of this model with the exponent -5/2 when a total ``mass'' has a certain critical value. Below the critical value, the system exhibits an ordinary critical behavior, and scaling relations hold. Above the threshold, numerical simulations show that a power-law distribution coexists with a huge ``monopolized'' member. It is argued that these behaviors are observed universally in conserved aggregation processes, by analizing an extended model.
△ Less
Submitted 22 October, 2007; v1 submitted 26 October, 2005;
originally announced October 2005.
-
Similarity and Probability Distribution Functions in Many-body Stochastic Processes with Multiplicative Interactions
Authors:
Akihiro Fujihara,
Toshiya Ohtsuki,
Hiroshi Yamamoto
Abstract:
Analytical and numerical studies on many-body stochastic processes with multiplicative interactions are reviewed. The method of moment relations is used to investigate effects of asymmetry and randomness in interactions. Probability distribution functions of the processes generally have similarity solutions with power-law tails. Growth rates of the system and power-law exponents of the tails are…
▽ More
Analytical and numerical studies on many-body stochastic processes with multiplicative interactions are reviewed. The method of moment relations is used to investigate effects of asymmetry and randomness in interactions. Probability distribution functions of the processes generally have similarity solutions with power-law tails. Growth rates of the system and power-law exponents of the tails are determined via transcendental equations. Good agreement is achieved between analytical calculations and Monte Carlo simulations.
△ Less
Submitted 25 August, 2005;
originally announced August 2005.
-
Power-law tails in nonstationary stochastic processes with asymmetrically multiplicative interactions
Authors:
Akihiro Fujihara,
Toshiya Ohtsuki,
Hiroshi Yamamoto
Abstract:
We consider stochastic processes where randomly chosen particles with positive quantities x, y (> 0) interact and exchange the quantities asymmetrically by the rule x' = c{(1-a) x + b y}, y' = d{a x + (1-b) y} (x \ge y), where (0 \le) a, b (\le 1) and c, d (> 0) are interaction parameters. Noninteger power-law tails in the probability distribution function of scaled quantities are analyzed in a…
▽ More
We consider stochastic processes where randomly chosen particles with positive quantities x, y (> 0) interact and exchange the quantities asymmetrically by the rule x' = c{(1-a) x + b y}, y' = d{a x + (1-b) y} (x \ge y), where (0 \le) a, b (\le 1) and c, d (> 0) are interaction parameters. Noninteger power-law tails in the probability distribution function of scaled quantities are analyzed in a similar way as in inelastic Maxwell models. A transcendental equation to determine the growth rate γof the processes and the exponent s of the tails is derived formally from moment equations in Fourier space. In the case c=d or a+b=1 (a \neq 0, 1), the first-order moment equation admits a closed form solution and γand s are calculated analytically from the transcendental equation. It becomes evident that at c=d, exchange rate b of small quantities is irrelevant to power-law tails. In the case c \neq d and a+b \neq 1, a closed form solution of the first-order moment equation cannot be obtained because of asymmetry of interactions. However, the moment equation for a singular term formally forms a closed solution and possibility for the presence of power-law tails is shown. Continuity of the exponent s with respect to parameters a, b, c, d is discussed. Then numerical simulations are carried out and campared with the theory. Good agreement is achieved for both γand s.
△ Less
Submitted 30 June, 2005;
originally announced June 2005.
-
Two-component sandpile model : self-organized criticality of the second kind
Authors:
Akihiro Fujihara,
Toshiya Ohtsuki,
Teruhiro Nakagawa
Abstract:
Two-component sandpile models are investigated numerically and theoretically. Monte Calro simulations are performed to show that probability distribution functions of avalanche size and lifetime obey power laws whose exponents are approximately equal to 1.5 and 2.0 and the system exhibits SOC. A mean-field theory is developed to discuss the essence of the processes. We find that two-component mo…
▽ More
Two-component sandpile models are investigated numerically and theoretically. Monte Calro simulations are performed to show that probability distribution functions of avalanche size and lifetime obey power laws whose exponents are approximately equal to 1.5 and 2.0 and the system exhibits SOC. A mean-field theory is developed to discuss the essence of the processes. We find that two-component models approach a steady critical state belonging to a different universality class from that of one-component models. Conservation of two kinds of sands at local toppling causes an infinite number of stable states which substitute for artificial boundary dissipation. Among two control parameters appearing in one-component models, therefore, a rate constant of dissipation is removed in two-component models. It is concluded that the more conserved quantities result in the less control parameters and a novel class of SOC.
△ Less
Submitted 11 July, 2005; v1 submitted 29 June, 2005;
originally announced June 2005.
-
Effects of Randomness on Power Law Tails in Multiplicatively Interacting Stochastic Processes
Authors:
Toshiya Ohtsuki,
Akihiro Fujihara,
Hiroshi Yamamoto
Abstract:
Effects of randomness on non-integer power law tails in multiplicatively interacting stochastic processes are investigated theoretically. Generally, randomness causes decrease of the exponent of tails and the growth rate of processes. Explicit calculations are performed for two examples: uniformly distributed and two peaked systems. Significant influence is demonstrated when a bare growth rate i…
▽ More
Effects of randomness on non-integer power law tails in multiplicatively interacting stochastic processes are investigated theoretically. Generally, randomness causes decrease of the exponent of tails and the growth rate of processes. Explicit calculations are performed for two examples: uniformly distributed and two peaked systems. Significant influence is demonstrated when a bare growth rate is low and coupling is weak. It should be emphasized that even the sign of the growth rate can be changed from positive to negative growth.
△ Less
Submitted 15 December, 2003;
originally announced December 2003.
