| When fitting complex models, such as finite element or discrete event simulations, the experiment design should exhibit good properties of both projectivity and orthogonality. To reduce experimental effort, sequential design strategies allow experimenters to collect data only until some measure of prediction precision is reached.;In Chapter 2, we present a batch sequential experiment design method that uses sliced Full Factorial-Based Latin Hypercube Designs (sFFLHDs), which are an extension to the concept of sliced Orthogonal Array-Based Latin Hypercube Designs (OALHDs). At all stages of the sequential design, good univariate stratification is achieved. The structure of the FFLHDs also tends to produce uniformity in higher dimensions, especially at certain stages of the design. We show that our batch sequential design approach has good sampling and fitting qualities through both empirical studies and theoretical arguments.;Software testing is often done in an ad hoc manner with various expert users simply trying many different common functions to find and report bugs in the software. In Chapter 3, we focus on a particular software structure that is common to options menus and other settings menus. The key aspect of this type of menu system is that the no real action is taken except at the very last level of the menu where some option is selected. That selected option then remains in force until a different option for the same function is selected. We provide a batch sequential design methodology that allows operators to continue to work for as long as they are able and continue to make progress on finding the most important errors first (presuming that all combinations will never be able to be tested). We focus on discovery of three types of errors: (1) Errors in the menu system that prevent the user from moving to the next level of the menu or inability to choose a certain option, (2) Errors in the performance of the code due to the selection of a certain option, and (3) Errors in the performance of the code due to the selection of a certain set of options. Since all possible sets of options are unlikely to be able to be tested, we begin with testing all possible pairs of option settings. |
