Math analyst at Lockheed, Memorex Corp, and others for 10+ years. Turned several simulation problems into optimization problems and reduced design times by many orders of magnitude.
Found answers to time series analysis using several spectral estimation methods and compared their results. FFT was dropped do to my work showing the FFT provided the poorest results.
Since 1975, I've have been teaching/training engineers & scientists on Calculus-level programming in order to improve their productivity. Calculus programming can solve most math problems in hours. It can tweak 100s of parameters to find an optimal solution. Its great.
In response to your 'Complete simulations' interest: I'm involved with a project that will examine ship survivability in terms of physical damage to systems, injuries inflicted on personnel located in affected compartments based on enemy weapon, angle of attack, location of attack, blast modeling of structure and compartment shape, injuries inflicted on crew, function of crew, treatment and outcome of injured, redundancy of crew in specific locations etc...the idea is to examine all possible (or enough to inform decision makers) system configurations with all possible enemy attacks/results from attacks to design a ship that can maintain mission capability with greatest probability
This seems to fall right in line with the type of problems you're interested. I'm also looking at designing casualty evacuation vehicles from a systems point of view (speed, range, capacity etc).