« Smart Mobs: The Net, the Dean campaign, and Technological Determinism | Main | The Selfish Class »

februari 04, 2004

Undersökning av Java-system för agentbaserad modellering

I dokumentet Evaluation of free Java-libraries for social-scientific agent based simulation, skrivet av Robert Tobias och Carole Hofmann, görs en utvärdering av Java-system för agentbaserad modellering. Dokumentet innehåller inte så mycket mer än en diskussioner kring och utvärdering av kriterier som sådana system bör ha.

Vissa krav uppfylldes dåligt av de testade system, t.ex. "Support for project organization" och "Support for communication". Tyvärr får kriteriet "Ease of Use" som bäst 3 av maximalt 6 poäng.

Det system som fick flest poäng är RePast, vilket också råkar vara denne bloggantecknares favorit. De övriga inblandade var Javaversionen av Swarm, Quicksilver (det verkar dock vara något konstigt med denna Sourceforge-länk; sök på "Quicksilver" och välj det projekt som är om agenter) samt VSEit. I appendix nämns de system som valdes bort och av vilken avledning.


Abstract:
This paper compares four freely available programming libraries for support of social scientific agent based computer simulation: RePast, Swarm, Quicksilver, and VSEit. Our aim is evaluation to determine the simulation framework that is the best suited for theory and data based modeling of social interventions, such as information campaigns. Our first step consisted in an Internet search for programming libraries and the selection of suitable candidates for detailed evaluation on the basis of 'knock out' criteria. Next, we developed a rating system and assessed the selected simulation environments on the basis of the rating criteria. The evaluation was based on official program documentation, statements by developers and users, and the experiences and impressions of the evaluators. The evaluation results showed the RePast environment to be the clear winner. In a further step, the evaluation results were weighted according to effort/time/energy saved by social scientists by using the particular ready-made programming library as compared to doing their own programming. Once again, the weighted results show RePast to win out over the other Java based programming libraries examined.

Summering av bedömningarna:

Even though the simulation frameworks evaluated show only minor differences in their scores on most of the individual criteria, taking all the criteria together yields a clear winner: RePast. The total scores of the lesser-known simulation frameworks, Quicksilver and VSEit, are similar and much lower than the total score of RePast. For 19 criteria and with a difference of at least 20 points, RePast is superior to these on average in all of these criteria. Quicksilver rates higher than RePast only on support for Monte Carlo simulations and for close contact between developer and users. What is surprising is that RePast beats Swarm, which was used as a reference here even though it is actually not suitable due to its programming language. Except for the criterion "user base," RePast equals or surpasses Swarm on every criterion.

Man har sedan en brasklapp angående värdet av den gjorda bedömningen:

The difficulty in evaluating these simulation frameworks lies mainly in the fact that it is practically impossible to gain extensive practical experience with the frameworks within a period of a few weeks. To be able to make valid statements about some key criteria, one would need at least several months of application practice. For this reason, the present evaluation is really mainly a theoretical one. The ratings were made mainly based on documentation and technical references. Where possible, we included information supplied by developers and users. It is very possible that the rating results would look very different after having implemented the frameworks over a long period of time.


P.S. Dokumentet är publicerat i tidskriften JASSS (Journal of Artificial Societies and Social Simulation). Där finns många andra relevanta dokument kring simulering av sociala fenomen etc.

Posted by hakank at februari 4, 2004 04:18 EM Posted to Agentbaserad modellering