SHELscape: Simulation Hub for Economic Landscapes, is a complex systems framework developed to understand patterns of post-natural disaster (earthquakes, floods, tusnamis) labor and goods movements and their impact on income and consumption distributions. The motivation for this research comes from a lack of baseline data in low-income regions where natural disasters usually affect a large population and with limited resources policy response, becomes a key challenge.

The model uses a bottom-up approach to define interactions at the micro agent-agent Level. The economy is programmed as primarily an rural agrarian society with multiple villages and cities. Each location reaches it own stable equilibrium in the goods and the labor markets which determines price and income/consumption distributions. Locations interact with each other through migration and regional export of goods resulting in meso-level stable trends at the regional level.

The system can be shocked through various channels including disruption of food production, loss of human life, disruption of road networks, and productivity loss. Agents adapt to the changing environment which allows for tracking regional level patterns of population displacement, and income and consumption distributions and identifying clusters of vulnerability.

Regular project updates can be followed here:

The presentation below gives a quick introduction to the baseline model setup:

The baseline version of the model is 2x2x2 economy with location types (villages, cities), agent types (owners, workers), and goods types (consumable goods, services). Locations are integrated with each other through road networks.

The interactions are defined by 4 micro modules:

1. Production
2. Wages
3. Buy
4. Consume

and 2 meso modules:
1. Goods market
2. Labor market




The model introduces several novel features in the use of ABM modeling;
1. Spatially defined locations
2. Multiple markets (labor, goods)
3. A* graph search algorithm for migration and selling decisions
4. Iterative search algorithm for selecting locations for selling goods
5. Probabilistic decision making procedures that allows for controlling randomness in outcomes


Naqvi, A. (2017). Deep Impact: Geo-Simulations as a Policy Toolkit for Natural Disasters. World Development, DOI: 10.1016/j.worlddev.2017.05.015.

Naqvi, A. and Rehm, M. (2014). A Multi-Agent Model of a Low Income Economy: Simulating the Distributional Effects of Natural Disasters. Journal of Economic Interaction and Coordination, 9(2) 257-309. DOI: 10.1007/s11403-014-0137-1

Naqvi, A. and Rehm, M. (2014). Simulating Natural Disasters – A Complex Systems Framework. 2014 IEEE Conference Proceedings on Computational Intelligence for Financial Engineering & Economics (CIFEr), 27-28 March, London, UK. DOI: 10.1109/CIFEr.2014.6924103.

Naqvi, A., Sobiech, C. (2010). Simulating Humanitarian Crisis and Socio-Economic Vulnerability: Applications of Agent-Based Models. United Nations University – Institute for Environment and Human Security SOURCE 13.

Naqvi, Ali Asjad (2010). Floods and Policy Planning. LUMS Social Science and Policy Bulletin.

Conference Presentations
ERSA GfR Conference, (Spital am Phyrn, Austria), February 2017
IIASA Risk, Policy, Vulnerability (RPV) Seminar series (Laxenburg, Austria), September 2014
LUMS Brown Bag Lecture series, (Lahore, Pakistan), May 2014
IEEE Conference on Computation Economics and Financial Research, March 2014
IIASA-UNISDR Conference (Laxenburg, Austria), June 2012
Crisismappers Webinar series, Jan 2012
FMM Conference (Berlin, Germany), October 2011
Eastern Economic Association Conference (NY, USA), June 2011
George Mason University, Department of Computational Social Science seminar series (Washington, USA),Feb 2011
NICO Complexity Conference, Northwestern University (Chicago, USA), August 2009
Eastern Economic Association Conference (NY, USA), March 2009

Leave a Reply

Your email address will not be published.