[R] Optimization when only binary variables can be manipulated?

From: Ben Fairbank <BEN_at_ssanet.com>
Date: Fri, 29 Feb 2008 10:42:40 -0600

I am trying to optimize in situations such as the following:  

Given 100 ability test items with such known item values as (1) difficulty, (2) correlation with criterion, (3) position in subject matter taxonomy, (4) illustrated/nonillustrated, (5) abstraction level, and (6) length, I seek to make three 20-item tests that are as nearly identical in their properties (difficulty, illustrations, taxonomy, etc) as possible, using each item only once. (The goal is to make the tests interchangeable; there are approx 2.6 e50 such sets of tests.) I have an expression for the merit of the extent to which the tests are identical, but since all of the manipulated variables are binary (i.e., each item is "in" or "out" of each of the three tests), derivative-based methods seem not to apply.  

I have read through the optimization chapter in MASS, but those methods appear not to cover this situation. Can any of the R optimization packages handle optimization when the manipulated variables are binary and numerous?  

With thanks for any suggestions,  

Ben Fairbank

Technical Director

Sinclair Customer Metrics


        [[alternative HTML version deleted]]

R-help_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Fri 29 Feb 2008 - 16:55:57 GMT

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.2.0, at Fri 29 Feb 2008 - 17:30:18 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.

list of date sections of archive