The circular Sisyphus random walk model

Shahar Hod
DOI: https://doi.org/10.1140/epjp/s13360-023-04816-2
2024-01-11
The European Physical Journal Plus
Abstract:We study, using analytical techniques, the time-dependent behavior of the circular Sisyphus random walk model, an infinite stochastic Markov chain whose dynamics is determined by the following jumping rule: at every discrete time step the walker has a finite probability q to move one step upward toward her target at the top of the 'ladder' and a complementary probability to fall back to her initial position, resulting in the need to restart the climbing process. A climber who reaches her target after making upward steps in a row (without a single stumble in between) is doomed to restart the circular Sisyphean process all over again from the bottom of the ladder. We determine the time-dependent success rate function of the system, the fraction of circular Sisyphus random climbers who reach their target after making exactly t steps. We identify a special family of initial conditions that leads to stationary configurations of the circular random climbers with constant (time-independent) success rate functions. Interestingly, it is explicitly proved that, for generic initial conditions, the characteristic relaxation time of the system toward an equilibrium steady-state configuration is independent of the size of the system and it diverges in the deterministic limit. The analytically derived results are used to analyze the underlying dynamics of the 'conditional success game', a highly risky investment strategy in which a reward , which is a monotonically increasing function of the height of the challenge, is offered to agents who succeed in making successful steps in a row.
physics, multidisciplinary
What problem does this paper attempt to address?