Cycling: An Appraisal of Convergence Patterns with Respect to Even and Odd N-Point Designs
Abstract: Cycling, a problem associated with the construction of D-optimal designs, slows down the convergence of a sequence to a desired optimum, whenever it occurs in a variance exchange process. In this paper, we study the pattern of convergence induced by cycling especially when the size, N, of the design measure, , is either even or odd. The linear, interactive and quadratic portions of a two-factor response function, are used in the investigation. Numerical illustrations are given to demonstrate that the pattern of convergence depends on N in the search for D-optimality. The computations were conducted in R version 4.1.1 (2021). The result shows that the pattern of convergence differs with respect to the nature of the N-point design; even or odd, and to the degree of the two-variable response function; linear, interactive, and quadratic.
Keywords: Cycling, Variance exchange process, D-optimality, Response function, Even and odd N-point designs
How to Cite?: Felix Nwobi, Okim Ikpan, "Cycling: An Appraisal of Convergence Patterns with Respect to Even and Odd N-Point Designs", Volume 12 Issue 6, June 2023, International Journal of Science and Research (IJSR), Pages: 2863-2870, https://www.ijsr.net/getabstract.php?paperid=SR23624193838, DOI: https://dx.doi.org/10.21275/SR23624193838