For office use only
T1
________________
T2
________________
T3
________________
T4
________________
Team Control Number
44348
Problem Chosen
F
For office use only
F1
________________
F2
________________
F3
________________
F4
________________
2016 MCM/ICM Summary Sheet
Towards A Hopeful Journey
The world is witnessing the largest refugee crisis since the horrors of World War II. Modeling refugee
immigration, which is crucial to tackle the problem, is an intricate issue that should embrace the
sophistication of social interrelated systems, and take into the consideration of refugee crisis on local
conditions. Concerning the structure of refugee crisis and the routes of migration, with the official data in
2014, we construct a refugee migration model and build a feedback system using network analysis
methodology and Cellular automaton to make precise simulation aiming to help figure out a set of efficient
policies.
In the first place, we establish a set of metrics to consider the determinant factors in refugee migrations so
that we define our measures and indexes, after which we set the start points of six given routes as six nodes,
and choose 14 countries where most refugees gather to be the nodes in Africa and Central East. With the
assumption that the refugees migrate nearer and nearer to Europe, we divide the nodes into 4 layers based
on the distances from node to node using cluster analysis. In that way, the refugees migrate within the
layers.
After that, we assume that refugees get limited information and initially build a random migration model
to determine the migrating factors between 2 nodes. And by Matlab simulation we get result indicating the
main routes reaching Europe, but the numerical data is inconsistent with real data. Hence, we adjust our
presumptions and revise our model.
Next, inspired by Gravity Model, we analyze the factors that affect the migration of refugees and integrate
them into a comprehensive attraction index. By collecting and calculating the statistics, we figure out the
weight between two nodes and the ratios of population distribution at the six start points
(0.07:0.101:0.41:0.369:0.05), whose correlation coefficient with real data R=0.98. So we go on with revised
model to unravel the optimal flow distribution under different conditions and the measurement of weight
from node to node backed by single-period simulation.
In addition, we expand the scale of nodes. Given that the feedback information of refugees in different
periods and the maximum capacity in each node, we simulate the migration progress of refugees with
Cellular Automaton and C++, and the ratios of 3 nodes are 0.0362:0.537:0.427, R=1. Therefore, we get to
know about the influences of government and non-government organizations on refugee migration.
As for the policy, we attach significant importance to the empathy that the receiving population of each
country must fit the present refugee condition. By scaling up our model, we find that the routes get
saturated except for routes in North Africa and Central East, which may trigger the detention of refugees
and eventually lead to illegal immigration. Meanwhile, we work out the relations between the stability and
capacity of nodes by building a Cobweb Model. We find that the refugee flows tend to be more and more
stable in nodes with bigger capacity. So we also propose to stabilize the flows, which contributes to better
resource allocation and aid from GO and NGO.
Finally, we test the sensitivity of our model and conclude the strengths and weakness. The model is quite
reliable in small scale but still needs advancement for larger and more precise simulation.