Knightian Decision Theory, Part II: Intertemporal Problems
Abstract
The theory of choice proposed in “Knightian Decision Theory, Part I” is here applied to intertemporal problems. An analogue of dynamic programming called maxmin programming is developed. Also, it is shown that detailed contingent planning may not be needed in order to achieve maximality, a program being maximal if no other program is preferred to it. In certain circumstances, a maximal program can be achieved by making a finite calculation in each period. This calculation ignores distant future states and could also ignore unlikely contingencies. A decision maker making such calculations would behave much like a satisficer.