Centro de Documentacion de Fundación MAPFRE

Notas: Sumario: This paper examines a competitive setting in which two agents invest in a risk-free and a risky asset while considering both their own wealth and their wealth gap relative to each other. With market parameters partially or fully unknown, the problem is formulated as a nonzero-sum differential game within a reinforcement learning framework. Each agent seeks to optimize a Choquet-regularized, time-inconsistent mean-variance objective. Using dynamic programming, the authors derive a time-consistent Nash equilibrium in an incomplete market. Under a Gaussian mean-return assumption, they obtain an explicit analytical solution that enables the construction of a practical reinforcement learning algorithm. The algorithm shows uniform convergence, despite the absence of a traditional policy-improvement guarantee, and numerical experiments confirm its robustness and effectivenessRegistros relacionados: En: Astin bulletin. - Belgium : ASTIN and AFIR Sections of the International Actuarial Association = ISSN 0515-0361. - 19/01/2026 Volume 56 Issue 1 - January 2026 , p. 154 - 180Materia / lugar / evento: Mercados financieros Gestión de riesgos Modelo estocástico Modelo Gaussiano Cálculo actuarial Matemática del seguro Modelos matemáticos Otros autores: Guo, Junyi
International Actuarial Association
Otras clasificaciones: 6

‌ Ver detalle del número

A Nonzero-sum game with reinforcement learning under mean-variance framework