Note—A Computational Refinement for Discrete-Valued Dynamic Programs with Convex Functions

A search procedure is presented for discrete-valued dynamic programming problems with the objective of minimising the sum of convex functions. This procedure may be more efficient than available methods if the sum is to be minimized over a uniform grid of values of the state variable. Furthermore, the procedure can be used to reduce the computational effort if only some of the return functions have a special structure.

• 저널Management Science · 23(11) · 1251–1255
• 토픽Optimization and Mathematical Programming · Control and Systems Engineering
• DOI10.1287/mnsc.23.11.1251
• 저자Don R. Robinson