I can see from your profile that you are interested in Climate Change. You might find this paper by Berger, T. and Troost, C. relevant.

Berger, T. and Troost, C.  (2014), Agent-based Modelling of Climate Adaptation and Mitigation Options in Agriculture. J Agric Econ, 65: 323–348. doi:10.1111/1477-9552.12045

The authors used Agent-Based Simulation (ABS) for supporting the implementation of  climate-related policies in agriculture by modeling land-use change and energy markets. 

I have to say that I read about ABS and attended a seminar about it but I have never applied myself. For what I know ABS is usually proposed to support decision making by modeling behaviors of agents and several scenarios. What do you mean by "real example"? Do you wish to have a case where ABS was implemented for solving a problem of an organization and evidences of the problem being solved are disclosed? 

Thank you Ilaria.  Ideally, yes, as you suggest it would be great to see a full example of use from modelling the situation, to informing a decision, and the outcomes and any evaluation.  But I would also be happy with examples where it is clear that the output of a specified model in a configuration intended to represent a real-world situation has been used as one of perhaps many inputs to a decision.  For example a traffic-modelling program used to decide the optimal sequencing for traffic lights in some city, or a crowd-modelling program used to design stadium entry and exit procedures.

We used a simple ABS of evacuation in order to plan tsunami escape routes for the city of Ramat-Gan, Israel. This was about a year ago.

You can see a short clip that demonstrates the model (captions are in Hebrew but I guess the message can still be understood): https://www.youtube.com/watch?v=gTIipel2dn8

https://www.youtube.com/watch?v=gTIipel2dn8

Thank you Yonatan, this is just the sort of example I was looking for.  Is there a journal paper or a description of the work somewhere?  (And are the tsunami escape routes you planned now signposted in the city?)

Thanks, unfortunately, we cannot publish the results.

The routes are supposed to be signposted, I don’t know exactly when. They do this now in all of Israel: http://www.timesofisrael.com/emergency-services-drill-tsunami-along-southern-coast/

There's an older thread here in ResearchGate (to which I tried to contribute) that you might find interesting.

https://www.researchgate.net/post/Are_there_any_examples_where_ABM_S_has_been_USED_pracically_by_stakeholders_to_support_decision_making_in_Operations_Research

Erica, 

Through my research I developed a disaster recovery decision making framework utilizing agent based modeling (ABM) in order to assimilate the needs of the various participating entities in the redevelopment activities. Such framework aims to increase the individual utilities of the different stakeholders and decrease the built environment vulnerability to future shocks. 

Below is a link to one of the published journal papers:

https://www.researchgate.net/publication/307553658_Sustainable_Disaster_Recovery_Decision-Making_Support_Tool_Integrating_Economic_Vulnerability_into_the_Objective_Functions_of_the_Associated_Stakeholders

Hope this helps

Mohamed 

Article Sustainable Disaster Recovery Decision-Making Support Tool: ...

Hello, Erica. I am interested in agent based modeling and its use in decision making in reality. My team study large scale epidemic spread in Beijing and emergency evacuation using agent based modeling. From my point of view, agent based modeling is a quantitative method, but its result is qualitative.

Looking better at Yonatan answer, I can also report of a work on preparedness and resilience against disasters, in particular in the city of Padang, in case of a Tsunami. Here's a link to a demonstration of the simulation results, and a link to the profile of a researcher I know working in the project.

http://matsim.org/scenario/padang-indonesia

https://www.researchgate.net/profile/Gregor_Laemmel

 Thanks Giuseppe, that's another interesting one.  Perhaps the human-agent simulations are easiest to verify against intuition ("What would I do in this circumstance?"), making them an easier sell to decision-makers.

Liang Ma, I like your observation that the method is quantitative but the result is qualitative.  Can you say a bit more about what that means in practice?  How do you use the result of your simulations?

In a way, simulations (and computer models) are prostheses of human thought for elaborating large "what-if" scenarios. Of course models must have a minimum level of proved reliability and accuracy in relevant scenarios, otherwise we are making bad video games. Moreover, consider that (at least in the transportation area) sometimes you get counter intuitive results (e.g. adding an obstacle can make pedestrian flow smoother)... so it's complicated, but the idea of providing decision-makers ways to test hypothesis is generally the way to introduce this kind of approach into a new context.

You might want to have a look at http://www.anylogic.com. There, they do provide references for Business/System Dynamics, Multi-Agent, and Discrete Event Simulations. 

Then, there is the field of traffic simulations. PTVs Vissim (Simulation of Traffic in Cities) is based on a car-following model, i.e. an individual model. For pedestrian and evacuation dynamics, there is a number of commercial software simulating individual agents (http://ped.wikia.com/wiki/Crowd_Models). The same holds for MATSim (Multi-Agent Transport Simulation). Senozon (http://www.senozon.com/references), a swiss "MATSim spin-off" is doing large scale traffic simulations, e.g. for Switzerland and now for Germany and are providing services (sort of geomarketing) based on that. Guiseppe has already mentioned the MATSim evacuation example. There is more in the MATSim book (http://matsim.org/the-book). Of course, the question is often, whether the same results could be obtained by simpler methods (e.g. spreadsheet calculations based on road capacities).