DOE Of War Games

From Mckwiki

Jump to: navigation, search

1 Summary

Ship acquisition decisions are made based on technical information provided by models of the ships under consideration - we don't actually build them and then choose, instead we rely on engineering depictions and predictions. In the ideal case these models are actually used in scenario evalution tools such as war games (for warships) or comprehensive economic models (for mer- chant ships.) It is the output from these scenario evaluation tools that then drives or informs the decision-making.

As ship design engineers we (for I am member of that community) spend a fair amount of effort to ensure that our ship model - our engineering depiction of the ships under consideration - is detailed and accurate. And this community has an active dialogue underway regarding what the appropriate level of detail and precision is, in consideration of which stage of acquisition is underway. One of the tools being used in ship design is the mathematics of Design of Experiments wherein technical principles are applied to determine the needed level of detail of an investigation, or the needed 'spacing' of design variables, and so forth. But to my knowledge no similar application of Design of Experiments has been made to the war game.

Do we know what variables the war game is sensitive to? Do we know that the outcome of the war game actually depends significantly on the characteristics of the ship? Or upon which characteristics of the ship does it depend? If we knew this, then we might be able to better deploy the ship design engineering effort so that we spend our time generating detail on the important parameters of the ship (as it were, perhaps survivability and maneuverability) rather than upon those characteristics that are more familiar to naval architects (such as speed/power performance or weight estimating.) It might be that a design that is +/- 20% in speed has no effect upon the outcome, while +1% survivability turns out to be tie-breaking.

Personal tools