Uncertainty and the design of building subsystems—A dynamic programming approach

This paper discusses the general problem of designing building subsystems which include probabilistic demands thus producing a stochastic problem. This class of problem, if linear, can be readily solved using classical stochastic linear programming. It is shown that the problem may be formulated as a stochastic dynamic program. A problem of designing the HVAC subsystem is formulated and solved as a dynamic program. Comparisons are made between the two methods. It is shown that dynamic programming is the more powerful method both in terms of obtaining the solution and in its post-optimality analysis. It is suggested that dynamic programming is a suitable technique for a wide range of building and architectural design problems.