Aristotle University Of Thessaloniki (AUTH)
Δυναμικές διεργασίες σε πολύπλοκα συστήματα με χρήση υπολογιστικών προσομοιώσεων
Abstract
dc:descriptionIn this thesis I use Monte Carlo simulations to study the properties of biased and unbiased random walks on complex networks. More specifically, reaction-diffusion processes such as the trapping process A+T-> T are studied, as well as the spreading of infection in the two species reaction-diffusion process A+B->2B. These models may be relevant in communication networks where data traverses the network packets, spread of a virus in networks of routers, social networks for rumor spreading etc. In addition, a biological population model was constructed to study the evolution of the rate of biological aging based on the phenotype of the individuals. For the trapping problem we develop a simple theory to account for the behavior of the survival probability in a variety of conditions. In Erdos-Renyi (ER) networks we find that the trapping process exhibits a non-exponential behavior which depends on both the number of traps and the size of the network. In SF networks, when the trap is placed in one the network hubs, we find a new scaling with the system size. We also examine in detail the formation of a depletion zone in the trapping reaction in networks, with a single perfect trap. We show that the depletion zone is absent in regular, ER, and SF networks. The particles are homogeneously distributed in regular and ER networks, with the depletion effect appearing in very sparse networks. We also investigate the efficiency of biased random walks in complex networks. We reveal a non-universal scaling of the MFPT with the system size N with no traps present. When the hub of an SF network fails, we find that even for a small bias, fg stays relatively constant with N, while for the unbiased walk it follows a power law scaling with the system size. We also study the dynamics of the infection of a two mobile species reaction, from a single infected agent in a population of healthy agents. Find the density of healthy particles ρ(t) to be an exponential function in the long time limit in 2D, 3D lattices, ER and SF networks. We also investigate the scaling of the crossover time tc from short to long time exponential behavior, which we find to be a power law in lattices and ER networks. This crossover is shown to be absent in SF networks, where we reveal the role of the connectivity of the network in the infection process. In addition, I introduce a simple model to study how natural selection acts upon aging, which focuses on the viability of each individual. It is able to reproduce the Gompertz law of mortality and can make predictions about the relation between the level of mutation rates (beneficial/deleterious/neutral), age at reproductive maturity and the degree of biological aging. The process of aging in this simple model is revealed to be fairly complex, yielding a rich variety of results.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2011
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kittas, Aristotelis
- Κίττας, Αριστοτέλης
Subjects
dc:subject × 14Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/26893