When Geometry Shapes Information in Pursuit-Evasion Games
Abstract
Information is typically treated as an exogenous component of multi-agent decision making, with agents adapting their behavior to observations generated by fixed sensing models. We challenge this assumption by introducing a geometry-aware pursuit-evasion framework in which sensing becomes an endogenous strategic variable. By coupling geometry-dependent observations, Bayesian belief updates, and stochastic Quantal Best-Response dynamics, we show that strategic motion reshapes the information available to opponents and deforms the resulting equilibrium structure. We characterize the limits of this information manipulation through Fisher information bounds, showing that sensing geometry constrains achievable belief distortion and the emergence of phantom regions. Experiments in continuous pursuit-evasion environments demonstrate that geometry variations reshape belief trajectories, modify equilibrium configurations, and reveal a fundamental tradeoff between physical evasion and information manipulation.
Related
- New strategies for some pursuit-evasion differential games with Gronwall-type constraints
- Analysis of First-Order Linear Pursuit and Evasion Differential Games with Gronwall-Type Constraints
- Convex Pursuit-Evasion Games for Spacecraft Proximity Operations on Circular and Elliptical Orbits
- Solution and Optimal Strategies for a Differential Game of Pursuit-Evasion with Point-Wise Constraints
- Dynamical Systems in Elliptical Pursuit and Evasion
BibTeX
@misc{cadre2026when,
title = {{When Geometry Shapes Information in Pursuit-Evasion Games}},
author = {Hélène Le Cadre and Ava Sedighi and Frédy Pokou},
howpublished = {HAL},
year = {2026},
month = sep,
url = {https://inria.hal.science/hal-05758551v2},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - When Geometry Shapes Information in Pursuit-Evasion Games
AU - Hélène Le Cadre
AU - Ava Sedighi
AU - Frédy Pokou
PY - 2026
DA - 2026/09/24
PB - HAL
UR - https://inria.hal.science/hal-05758551v2
AB - Information is typically treated as an exogenous component of multi-agent decision making, with agents adapting their behavior to observations generated by fixed sensing models. We challenge this assumption by introducing a geometry-aware pursuit-evasion framework in which sensing becomes an endogenous strategic variable. By coupling geometry-dependent observations, Bayesian belief updates, and stochastic Quantal Best-Response dynamics, we show that strategic motion reshapes the information available to opponents and deforms the resulting equilibrium structure. We characterize the limits of this information manipulation through Fisher information bounds, showing that sensing geometry constrains achievable belief distortion and the emergence of phantom regions. Experiments in continuous pursuit-evasion environments demonstrate that geometry variations reshape belief trajectories, modify equilibrium configurations, and reveal a fundamental tradeoff between physical evasion and information manipulation.
ER -