This repository contains a comprehensive decision support analysis for a seasonal ski rental business. The project evaluates the interplay between controllable variables (Pricing, Fleet Size) and uncertain environmental factors (Season Length, Daily Demand) using advanced simulation techniques.
The goal of this project is to move beyond "best-case" estimates and provide a robust probabilistic view of business performance. We identify the optimal "sweet spot" for pricing and fleet capacity while quantifying the risk of financial loss.
- Monte Carlo Simulation: 10,000+ iterations using Latin Hypercube (QMC) sampling to cover the entire uncertainty space.
- Global Sensitivity Analysis: Comparison of Sobol' Indices (variance-based) and Regional Sensitivity Analysis (Monte Carlo filtering) to isolate profit drivers.
- Scenario Decomposition (SimDec): Visual partitioning of profit distributions to identify specific "paths" to success.
- Risk Quantification: Detection of "tail risks," including a calculated 0.07% probability of seasonal loss.
- Pricing Strategy: The optimal rental price is identified between €70–€80. Prices below this range significantly increase the risk of insolvency during short seasons.
- Diminishing Returns: Due to price elasticity and capacity constraints, fleet expansion beyond 100-110 units yields diminishing returns without a guaranteed "long" season.
- Interaction Effects: Global sensitivity shows that while Price is the strongest lever, its effectiveness is physically bounded by Season Length, necessitating a conservative approach to fixed-cost expansion.
- Python (Pandas, NumPy, Matplotlib, Plotly)
- SALib (Sobol' Sensitivity Analysis)
- SimDec (Scenario Decomposition & Visualization)
- SciPy (Statistical testing and KS-statistics)
The model was created as a school assignment for the course System Dinamics with Applications.