This research aims to develop a unified framework for integrating surveys so as to guide the design of a multi-purpose survey program. Algorithms will be developed for optimal integration based on more than one objective function and for sequential integration of surveys. The investigators will attempt to use the one-to-one correspondence between general sampling schemes and unit-by-unit sequential sampling to reduce the size of the underlying transportation problem, thereby expediting the solution of the integration problem and, in certain cases, providing solutions in simple closed form. This research addresses the problem of integrating surveys. The problem of designing a cost-effective sampling program for two or more surveys which maximize the overlap between the old and the new samples is known to be of considerable practical interest in the design of large-scale periodic and multi-purpose surveys. Although this problem has been studied for several decades, no unified theory for it has yet been developed. The object of this project is to develop a unified framework within which to evaluate the integration of surveys. New algorithms and computer software will be developed for optimal integration of surveys under general sampling schemes. Optimality properties of these and other existing methods will be evaluated.
