Loading...

Modelling the Physics of Collective Decision Making and Commuication by Signal Exchange

Salahshour Mehmandoust Olia, Mohammad | 2019

666 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 52337 (04)
  4. University: Sharif University of Technology
  5. Department: Physics
  6. Advisor(s): Rouhani, Shahin; Roudi, Yaser
  7. Abstract:
  8. In many biological populations, information about environmental conditions is acquired collectively, through information sharing among individuals, by exchanging signals. In order to achieve a better understanding of such systems, using methods borrowed from statistical physics, we introduce two mathematical models. First, a collective decision making model, for a community of individuals residing on a communication network, who live in an uncertain environment, in which individuals try to collectively find the environmental state. In the second model, a collective movement model, we consider a population of collectively moving individuals, in which individuals try to collectively find and travel to an environmental direction. In both models, individuals are able to make noisy observation of the environment, and share their information by production and comprehension of signals. We show that in both models, as noise in communication increases, a transition from an ordered phase in which consensus is formed, to a disordered phase in which no consensus is formed occurs. In addition, in both models, the ordered phase is composed of two phases. For high fraction of informed individuals, an informed consensus phase occurs, in which the population finds the environmental state. As the fraction of the informed individuals decreases, a discontinuous phase transition to a misinformed consensus phase occurs in which the population forms a wrong belief about the environmental state. Importantly, by showing that an amount of noise in signal production is more detrimental than the same amount of noise in signal comprehension, we show that the model predicts a fundamental asymmetry between signal production and comprehension in biological communication. By extending the collective decision making model, we then proceed to develop a mean field theory for this model. Using the mean field theory and agent based simulations, we derive the full phase diagram of the system and show that the system can be found in different multi-stable or mono-stable states. Given the different states that the system can be found in, an important question is that under what condition the population reaches the highest information acquisition capabilities? By considering the population in a changing environment, we show that the inference capability of the population and its responsiveness to environmental changes is maximized on the the edge of bi-stability. That is on the border between the mono-stable informed consensus state and the bi-stable informed-misinformed consensus. We discuss how this optimality condition contrasts the criticality hypothesis, according to which many biological functions are maximized at a critical transition. We also show that an optimal level of noise in communication, by increasing the decision making speed in response to environmental changes, in a phenomenon which resembles stochastic resonance, is beneficial in a changing environment
  9. Keywords:
  10. Communications ; Collective Sensing ; Collective Decision Making ; Cooperative Movement ; Signal Comprehension-Production ; Comprehension-Production Asymmetry ; Optimal Information Use

 Digital Object List

 Bookmark

...see more