The speed of a biased random walk on a percolation cluster at high density (Q1958458)

The speed of a biased random walk on a percolation cluster on \({\mathbb Z}^d\), \(d \geq 2\) in function of percolation parameter \(p\) is studied. A percolation cluster on the set of edges of lattice \({\mathbb Z}^d\) is a random configuration \(\omega\) where each edge has probability \(p\) of being open and probability \(1-p\) of being closed independently of all other edges. Transitions are allowed through open edges only. Given a vector bias \(\ell\), random walk on a percolation cluster is a Markov chain with transition probabilities defined by: \(p^\omega(x,x)=1\) if \(x\) has no neighbor in \(\omega\) and \(p^\omega(x,y)=e^{(x+y)\cdot \ell}/\sum_{x \sim z} e^{(x+z)\cdot \ell}\) otherwise, where \(x \sim z\) means that \(x\) and \(z\) are neighbors in \(\omega\). A first order expansion of the speed at \(p=1\) is obtained which proves that percolating slows down the random walk, for example, in the following cases: i) the drift is along the component of the lattice, ii) the drift is sufficiently weak. The methods rely on usage of Kalikow's auxiliary random walk adapted to considered model.