Research in
Agent-Based
Modeling
with
Mike Long
Contact information
melsch@RIT.edu
(585) 475-5158
Carlson 76-3131
 

Agent Based Modeling - Some Future Projects


Agent Based Modeling of Taxation


Currently, most arguments on taxation revolve around the degree and method of taxation and associated spending and not on taxation itself.  There are essentially three methods of taxation: Progressive, Regressive and Proportional.  In turn, the taxation of society has four main purposes or effects: 
    • Revenue - raise money for projects to maintain of improve
       the social welfare of the society
    • Redistribution – transfer the wealth from the rich to the poor
    • Repricing – attempt to shape behavior and penalize
    • Representation – leads to accountability of government, although it has been argued that elected
       governments and politicians do not maximize the welfare of society but the probability of winning an election.

One might then consider for a complete, and well-structured modeled society of different configurations or at various levels of development, e.g., developing vs. developed society, agricultural vs. industrial, urban vs. rural, etc., what is the effect of different modes of taxation?  What is the effect of the level of taxation?  When does over-taxation occur and what  is the effect on various attributes of the society?  What is the effect of inefficiency of taxation on the society and at what  level does it have an impact?


Agent Based Modeling of Skier Flow & Congestion

Snow ski and snowboard areas often have problems with skier congestion at trail intersections and lifts.  This can become particularly difficult when advanced, high-speed skiers merge with beginner skiers. Ski areas expend considerable effort to control skier flow and reduce congestion, out-of-control skiers, and collisions.  Agent-Based Modeling may be a method to gain insight and control this adverse behavior.  The objective of the program is to investigate various approaches to reduce skier collisions and congestions.

A similar problem was modeled with the emergency evacuation of buildings {D. Helbing, I. Farkas and T. Viscek, “Simulating dynamical features of escape panic”, Nature, 407, 487-490 (2000)}.  It was found that bisecting the exit with an obstruction reduced crowding, impeded congestion and added in the smooth and continual evacuation of the building.  Obviously an obstruction in the middle of a ski slope is not applicable, but other similar solutions to the problems of skier congestion might exist.

A graph is a collection of points (called vertices) and lines (called edges) where two vertices are connected by an edge. These graphs provide a general framework for studying a diverse array of scenarios including the investigation of social networks. A problem of interest involves a coloring of the vertices with colors 1,2,3,...,k. where two vertices that are connected by an edge can not receive different colors, where the overall goal is to minimize the number of colors used.

These ideas can be used to model skier congestion. We identify each route on a ski slope with a vertex and connect two vertices with an edge if and only if the two routes cross at one or more points. We are investigating how various vertex colorings of this graph can be used to optimize skier flow while minimizing collisions.

An Agent-Based Model of the Influence of
Special Interest Groups and Lobbying

What is the relative influence of special interest lobbying on the US Congress, both domestic and foreign.  Agent-Based Mdoeling and network theory would be well suited for this type of complex problem with multi-facet agents consisting of citizens, members of legislative body, government employees, members of special interest groups, special interest groups, coalitions, and lobbyist all with heterogeneous properties and interests such as political affiliations, incomes, education, employment, interests, etc.