Monte Carlo simulation is a computational method that uses random sampling to estimate the statistical properties of complex systems. In earthquake engineering, it is used to propagate uncertainty through probabilistic seismic hazard and risk analyses, generating distributions of outcomes instead of single deterministic values.
Monte Carlo simulation is a computational method that uses repeated random sampling to estimate the statistical properties of complex systems. The method is named after the famous casino in Monaco, reflecting its reliance on randomness. The basic principle is simple: if a system's behavior depends on uncertain inputs (each with a known probability distribution), we can generate many random samples from those distributions, compute the system's response for each sample, and analyze the resulting distribution of responses. As the number of samples increases, the estimated distribution converges to the true distribution β with error decreasing as 1/βN, where N is the number of samples. The method was developed in the 1940s at Los Alamos as part of the Manhattan Project and has since become one of the most widely used computational techniques across science, engineering, finance, and beyond.
Monte Carlo simulation is central to probabilistic seismic hazard and risk analysis. In probabilistic seismic hazard analysis (PSHA), the hazard at a site is computed by integrating over all possible earthquake scenarios β each defined by magnitude, distance, and other parameters β weighted by their probability of occurrence. Monte Carlo simulation provides an alternative to the classical integration approach, allowing the hazard curve to be computed by directly simulating earthquakes and ground motions according to their probability distributions. In seismic risk analysis, Monte Carlo simulation is used to propagate uncertainty through each stage of the risk chain β from ground motion to structural response to damage to loss β generating distributions of repair cost, downtime, and casualties. This probabilistic framework, formalized in FEMA P-58 and the PEER PBEE methodology, has become the standard for performance-based earthquake engineering. Monte Carlo is also used in reliability analysis, where it estimates the probability of failure of a structural system under uncertain loads and resistances, and in model updating, where it samples from the posterior distribution of model parameters given measured data.
Monte Carlo simulation is powerful but computationally demanding. The accuracy of the method improves slowly with the number of samples (as 1/βN), so achieving high accuracy requires many samples β often 10,000 to 1,000,000 or more. For computationally expensive simulations (e.g., nonlinear response history analysis of complex structures), this can be prohibitive. Several techniques have been developed to improve efficiency. Variance reduction techniques β such as importance sampling, stratified sampling, and antithetic variates β reduce the number of samples required for a given accuracy. Latin hypercube sampling provides better coverage of the input space than pure random sampling. Surrogate models β also called metamodels, response surfaces, or emulators β replace the expensive simulation with a fast approximation, allowing many more samples to be evaluated. Subset simulation and line sampling are advanced techniques for estimating small failure probabilities that would be intractable with standard Monte Carlo. Markov Chain Monte Carlo (MCMC) is used for Bayesian inference, where the goal is to sample from a posterior distribution rather than to estimate a simple probability. In earthquake engineering, Monte Carlo simulation is used in PSHA, seismic risk assessment, reliability analysis, and increasingly in real-time applications where fast probabilistic estimates are needed. In Iran, Monte Carlo simulation is used in research and in seismic risk studies at regional and urban scales, following international standards such as those developed by GEM, HAZUS, and the PEER center.