In 1995, Dr. Paul Kaminski, then Defense Undersecretary for Defense for Acquisition and Technology, established the principles of CAIV-"Cost as an Independent Variable"-to take advantage of aerospace's experience using Target Cost to battle the escalating cost of defense systems in a time of declining procurement and development dollars. One of the key tenets of CAIV is to limit Key Performance Parameters (KPP) to the critical few, to establish thresholds and desired levels of performance, and to trade off performance, cost and schedule. This trade arena is called the "Trade Space". Although a number of authors have proposed methodologies for performance, cost and schedule trades, a clear process has not emerged. One difficulty that emerged from study was how to use existing cost estimating systems to perform "Trade Space" optimizations. This paper presents the use of a cost model derived from log linear regression analysis of data from eighty-six NASA projects, the Analytic Hierarchy process, and non-linear optimization to show how CAIV "Trade Space" can be developed from cost data.