Invited Talks

Prof. Dr. Armando Geller (Group W and George Mason University, USA)

 

Title: Putting social simulation to use for security professionals and aid and relief practitioners in areas affected by armed conflict

 

Abstract: Successful development aid programming depends on knowing the needs of communities to be assisted. Eliciting development needs of a community is riddled with difficulties even at peace time; let alone during armed conflict when contending factions within a community overlap their development needs with allegiances to warring parties, obscuring "true" community needs with the operational and political dictates of the day. A novel need elicitation procedure is reported used around Tirin Kot, the capital of Uruzgan in Southern Afghanistan, an area characterized by strong Pashtun traditions, geographic remoteness, and sustained levels of violence over the last two years. Consequently, local communities are distrustful and fearful, suffer from physical damage to life and property, and depend on a local aid industry propped up by the inflow of development funds from Western donors. Mindful of the limits of surveys, case studies, quantitative approaches, and heuristic frameworks such as Tactical Conflict Assessment and Planning Framework (TCAPF) and District Stability Framework (DSF), participatory simulation modeling is leveraged to devise a "development game" in which participants (local powerbrokers and landowners) allocate development aid projects to different communities and report their underlying reasoning afterward. Development games elicit allocation matrices across the participants and access their latent, heterogeneous motivations and incentive structures. Participants' project portfolios are analyzed to (a) assess development needs and conflicts induced by development aid, (b) recommend minimum-cost, equitable project portfolios for small and large projects across development need dimensions, and (c) glean the rationale behind participants' decision making. "Need surfaces" are also derived and the uncertainty associated with them is estimated. For example, participants allocated small projects almost uniformly across communities and large projects to predominantly their home villages. They also prefer large projects to small ones, because large projects ensure that even under some misuse of funds the poor almost always receive project benefits (trickles down effect). Next development portfolio designs are tackled that reduce community grievances by guaranteeing a minimum level of need satisfaction for all communities. Finally, development portfolios are studied that minimize the difference in need satisfaction between the best and worst-off communities. It is concluded with a discussion of the limitations of the approach and an outlook for further work.

 

Prof. Dr. Alan Kirman (GREQAM, AMU and EHESS Marseille, France)

 

Title: Individual and Collective Rationality in Complex Systems

Abstract: This talk will analyse the emergence of collective behaviour which has a certain rationality from the point of view of the aggregate but which does not correspond to that of the individual. In economics it has long been recognised that, as a result of “externalities” behaviour that is rational from the point of view of the individual may lead to collective disasters as in the case of the “Tragedy of the Commons”. However, even more interesting are the cases where individuals who interact locally and have only limited information can together produce outcomes which have regularities which could not have been predicted from the individual motives. These may be beneficial but may also, as a result of contagion, lead the system to sudden collapse. Modelling this sort of phenomena usually involves starting with a formal analysis of a very simple case and then simulating a more complex and realistic system to see whether the results for the simple case still hold in the more general case. I will illustrate this with a number of examples, from some observations about the analogies with social insects, and ant and bees in particular, to recent work on the extension of the Schelling model of segregation. In the latter we include preferences for neighbourhood income as well as racial composition and introduce a housing market. We analyse segregation in terms of both race and income and the relation between house prices and these factors. I will also discuss a model of the market for mortgage backed securities and show how the rational reactions of a few individuals to a small change in the probability of default on the underlying assets can lead to a collapse in the price of the assets. In each case the self organisation of a complex adaptive system produces aggregate results which could not have been predicted from the analysis of a “representative agent”.

Sobre:

O Centro de Ciências Computacionais da Universidade Federal do Rio Grande é uma das 13 Unidades Acadêmicas da FURG, sendo responsável pela área das Ciéncias Computacionais, dedicando-se a formação de recursos humanos e a produção de conhecimento.

logo do centro de ciências computacionais logo da universidade federal do rio grande