Gains and costs of information in stochastic programming

The paper reviews the role of sensors as a variable depicting information in a stochastic program. Through examples, and with mathematical analysis of random measures, it is shown how policy variables affect the sensors in a given problem. Sensors are amenable to mathematical derivations. It is shown how the trade-off between the gain from introducing new information into the problem, and the cost of that action, can be addressed. Suggestions for further investigation are offered.