Hands-On Prescriptive Analytics

Chapter 7. Simulation Essentials

I will introduce simulations in this chapter as a powerful tool for Prescriptive Analytics. The inherent uncertainty in all decisions is the reason for simulations. As I argued in Chapter 1, all decisions made today are for a return tomorrow, whether “tomorrow” is literally tomorrow, next week, next month, next quarter, next year, or next decade; the return is in the future. The actual future period is immaterial regarding the uncertainty associated with the decision because it will be there regardless.

A major feature of a return is that you cannot say what it will be. It is due to shocks or disturbances with unknown or unknowable causes. Jurado et al. (2015, p. 1177) note that “uncertainty is typically defined as the conditional volatility of a disturbance that is unforecastable from the perspective of economic agents.” In our context, the economic agents are the business decision-makers, but this situation holds for all decision-makers, whether they are in the business domain, the public policy domain, or just ordinary citizens deciding to buy a new house, invest in stocks, or accept a job offer.

A decision-maker, acting today, must factor uncertainty into the decision-making process. But uncertainty per se cannot be measured. There have been many attempts to develop proxies or indicators of uncertainty that can be measured and tracked over time. Jurado et al. (2015, p. 1178) note that these include:

• Volatility of stock market returns measured ...