Abstract The present paper is an attempt to derive the equation of the maximum principle for adjoint processes in the general optimization problem, which is characterized by a situation when the choice of the control determines the absolutely continuous change of some basic measure. In contrast to the case of diffusion Markov processes ([1], [2]), these equations, generally speaking, are not ordinary differential equations here. We base the derivation on a non-linear equation for the martingale component of the value process ([3]) and an expression for the differential of the maximum of the semi-martingales. The method can be also applied to the case with jump components in the martingales defining measure densities, but we shall restrict ourselves by the continuous case.
Stochastic maximum principle in the problem of optimal absolutely continuous change of measure
1986-01-01
10 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Absolutely Stochastic Simulated Annealing Approach to Large Scale Unit Commitment Problem
Online Contents | 2006
|