solidot新版网站常见问题,请点击这里查看。
消息
本文已被查看4618次
A Convex Optimization Approach to Dynamic Programming in Continuous State and Action Spaces. (arXiv:1810.03847v1 [math.OC])
来源于:arXiv
A convex optimization-based method is proposed to numerically solve dynamic
programs in continuous state and action spaces. This approach using a
discretization of the state space has the following salient features. First, by
introducing an auxiliary optimization variable that assigns the contribution of
each grid point, it does not require an interpolation in solving an associated
Bellman equation and constructing a control policy. Second, the proposed method
allows us to solve the Bellman equation with a desired level of precision via
convex programming in the case of linear systems and convex costs. We can also
construct a control policy of which performance converges to the optimum as the
grid resolution becomes finer in this case. Third, when a nonlinear
control-affine system is considered, the convex optimization approach provides
an approximate control policy with a provable suboptimality bound. Fourth, for
general cases, the proposed convex formulation of dynamic programming op 查看全文>>