Multitime stochastic maximum principle on curvilinear integral actions

TL;DR

This paper develops a novel multitime stochastic maximum principle framework using stochastic curvilinear integrals, expanding stochastic control theory to multitime systems with new calculus rules and integral actions.

Contribution

It introduces a new multitime stochastic maximum principle based on curvilinear integrals, with original results on stochastic calculus, integrable systems, and control constraints.

Findings

Formulated multitime maximum principle for stochastic control.

Developed stochastic calculus rules for curvilinear integrals.

Provided examples of path independent and volumetric processes.

Abstract

Based on stochastic curvilinear integrals in the Cairoli-Walsh sense and in the It\^{o}-Udri\c{s}te sense, we develop an original theory regarding the multitime stochastic differential systems. The first group of the original results refer to the complete integrable stochastic differential systems, the path independent stochastic curvilinear integral, the It\^{o}-Udri\c{s}te stochastic calculus rules, examples of path independent processes, and volumetric processes. The second group of original results include the multitime It\^{o}-Udri\c{s}te product formula, first stochastic integrals and adjoint multitime stochastic Pfaff systems. Thirdly, we formulate and we prove a multitime maximum principle for optimal control problems based on stochastic curvilinear integral actions subject to multitime It\^{o}-Udri\c{s}te process constraints. Our theory requires the Lagrangian and the Hamiltonian as stochastic 1-forms.