What is percolation?

Percolation theory is a mathematical field that studies the behavior of connected clusters in a random graph. It seeks to understand how connectivity emerges when parts of a system are randomly connected. It finds applications in various areas, including physics, materials science, epidemiology, and computer science.

At its core, percolation concerns itself with the following:

  • The Model: Usually, the system is represented by a lattice (like a grid) or a random graph. Each site (or edge in a graph) is independently and randomly assigned to be either "open" (allowing flow) or "closed" (blocking flow) with a certain probability, p. The parameter p is the <a href="https://www.wikiwhat.page/kavramlar/percolation%20threshold">percolation threshold</a>.

  • Clusters: Connected components of "open" sites form <a href="https://www.wikiwhat.page/kavramlar/percolation%20cluster">percolation clusters</a>. These are groups of connected open sites.

  • Percolation: A fundamental question is whether an infinitely large cluster exists that spans the entire system. If such a cluster exists, we say that <a href="https://www.wikiwhat.page/kavramlar/percolation">percolation</a> occurs.

Key Concepts:

  • <a href="https://www.wikiwhat.page/kavramlar/Critical%20Exponents">Critical Exponents</a>: Near the percolation threshold, certain quantities like the size of the largest cluster and the correlation length exhibit power-law behavior. The exponents that describe these power laws are called critical exponents. These exponents are universal, meaning they depend only on the dimensionality of the system and not on the specific details of the lattice.

  • Universality: Systems with different microscopic details can exhibit the same critical behavior near their percolation thresholds. This is known as universality.

  • Applications: Percolation has real-world applications in modelling the flow of fluids through porous media, the spread of diseases, the electrical conductivity of disordered materials, and the resilience of networks.