Simulation Decomposition: New Approach for Better Simulation Analysis of Multi-Variable Investment Projects

This was the first published appearance of what later became SimDec.

At the time, it was not a “method.” It was a practical trick developed during my PhD research, supported by my supervisors, to better understand simulation results in a complex investment model. We were working on renewable energy policy analysis, running Monte Carlo simulations, and staring at histograms that clearly contained structure, but didn’t explain themselves.

The idea was simple: instead of accepting a single blended output distribution, classify simulation runs into meaningful managerial states (e.g., high vs. low capacity factor, localization fulfilled or not) and decompose the histogram accordingly.

That was it.

No grand claims. No quantification of variable importance. No unification of sensitivity analysis and uncertainty analysis yet. Just a structured way to see which combinations of variables actually drove profitability.

Looking back, this paper shows SimDec in its most primitive form: a single-case enhancement to simulation analysis. The broader methodological vision came later. But the core intuition — that one simulation contains many actionable insights — was already there.