Isotropic component properties are advantageous for some ballistic applications. Thus, given the high strength and refractory nature of Mo, powder metallurgy is the preferred processing approach for producing Mo components with the desired isotropic qualities. Ideally, proper process design allows for the production of parts with a specific relative density and degree of grain growth and a minimum of processing time, temperature, pressure, and material waste. However, the development of accurate models that provide a sufficient level of predictability can require expensive and time-consuming experimentation, which also is a factor in the overall waste stream and cost of production. Consequently, there exists a need to quickly evolve accurate physics based computer models with a minimum of experimentation. This goal has been shown to be achievable when model development and experimental design are allowed to co-evolve. One approach to this co-evolution is through the use of genetic algorithms and Bayesian inference.