Abstract

Previous reports on "Monte Carlo in Statistics" studies (ours and others') present tutorial and research fronts. Assuming early (book length) references represent culminations of work preceding their published dates by (a few) years, we can relate the efforts back to the beginnings of the computing era. The tutorial work includes treatment of statistical theory at least at an elementary level and it and the work reported in this paper serve to set the stage for higher-level probes, e.g., into more advanced and specialized topics such as the general linear model and (aspects of) multivariate statistics. Selected techniques among these latter topics are useful in providing service to simulators and putatively set the stage for collaborative computing, say, within an agent or grid purview. Our comments on these matters are based on some of our recent activities. Integrating statistical theory and practical computation in one scheme puts the entire statistical enterprise into a computation framework from which advanced thinking can be put to the test, e.g., whether a given test is adequate to its assignment or whether one approach is to be preferred over another. We briefly address some of these topics too.

Key Words

Monte Carlo techniques, agent systems, bio-computation, computer system simulations