Maybe Geosimulation software will be one of your potential choices, see e.g. http://www.geosimulation.org/

... Especially - related to your problem domain, they are building earthquake models: http://www.geosimulation.org/earthquake.htm

I have used @Risk to perform a Monte Carlo simulation. My first comprehensive use included my doctoral dissertation. While we were not modeling evacuation in the aftermath of an earthquake, we were modeling patient distribution based on theoretical availability at hospitals across a region of the US involved in burn care. To make it work, you'll need to identify your inputs; determine the type of probability distribution curve, known and potential variables to test your hypothesis, and calculate your coefficient of variation as an input. From there, plug in your inputs, step back and let the model run. Anyway, Monte Carlo simulation can give you the variety of probabilities from the obscure to the most likely. I'm sure there are other options and Monte Carlo simulation can be challenging but it's a great approach for most problems with multiple solutions that you want to measure the outcomes to include cumulative probability.

http://www.palisade.com/risk/

Thanks Randy and Mario

Dear Umma,

both geosimulation and @RISK seems to be very interesting tools. However, I think that the actual evacuation models denote many problems in simulation, expecially for their poor bases about human behaviors. The definition of model should be founded on experimental analysis of real cases. At the same time, the agent-based techniques (ABM) are able to effectively define the theoretical and operative simulation models for their rules-based aspects. This is very important especially in complex evacuation simulations, where people desires are not only connected to motion phenomena. Our research group was recently involved in a similar problem, and we provided a tool for simulating an earthquake evacuation in urban scenarios: EPES. This is based on the analysis of real world evacuations. EPES offers a combination between ABM for describing people desires and the Social Force model for describing people motion. EPES was validated through a cross-comparison with real evacuation data, and now we are trying to improve it from an operational point of view! I hope my answer will be useful for your question!

Article EPES– Earthquake pedestrians׳ evacuation simulator: A tool f...