Yunus Emre Demirci, Somnath Pradhan, Serdar Yüksel · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.28377 · cited by 0 works (OpenAlex) · open-access PDF
Nossas principais contribuições são as seguintes: (i) Mostramos a quase-optimalidade, sob o horizonte finito, bem como os critérios de custo do horizonte infinito descontados, de políticas de controle contínuo obtidas por meio de soluções para as aproximações do processo de decisão de Markov (POMDP) parcialmente observáveis em tempo discreto, construídas por meio de aproximações de Euler-Maruyama de ambos os processos de observação; (ii) obtemos várias propriedades de regularidade de tais modelos aproximados de POMDP; e (iii) usando essas propriedades de regularidade, chegamos a aproximações de modelo finito ideais. (iv) Estes culminam em nosso resultado principal, que é que uma solução ideal para o modelo finito aproximado de POMDP é perto do ideal para a difusão original
Machine translation (Meta M2M100); the English original follows.
We study continuous-time stochastic control problems under partial information. Our main contributions are the following: (i) We show near optimality, under finite horizon as well as discounted infinite horizon cost criteria, of piece-wise constant control policies obtained via solutions to discrete-time partially observable Markov decision process (POMDP) approximations, constructed via Euler-Maruyama approximations of both state and observation processes; (ii) we obtain several regularity properties of such POMDP approximate models; and (iii) using these regularity properties, we arrive at near optimal finite model approximations. (iv) These culminate into our main result which is that an optimal solution for the approximate finite POMDP model is near optimal for the original diffusion with partial information.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Partially observable Markov decision process, Optimal control, Mathematics, Observable, Markov process, Applied mathematics
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.28377, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange