Home

OSTRICH is a model-independent program that automates the processes of model calibration and design optimization without requiring the user to write any additional software. Typically, users only need to fill out a few required portions of the OSTRICH input file (i.e. ostIn.txt) and create template model input files. Users may also activate and configure a variety of optional features, including: parallel processing, model pre-emption, algorithm restarts, parameter statistics and regression diagnostics, telescoping parameter bounds, predictive uncertainty, non-ASCII model I/O, and user-defined initial parameter sets. Finally, users with skill in code or script development (e.g. C/C++, FORTRAN, Java, R, MATLAB, Python, bash/bat, etc.) may also wish to take advantage of OSTRICH’s parameter correction feature. It allows users to correct candidate parameter sets based on rules-of-thumb or expert judgment that would otherwise be difficult to encapsulate as optimization constraints.

The remainder of this manual describes the configuration and usage of OSTRICH. Sections 1.1 and 1.2 provide brief summaries of the currently supported optimization and calibration algorithms (Section 1.1), and regression statistics and diagnostics (Section 1.2). Sections 2 through 2.24 describe the OSTRICH input file and its various configuration sections. The majority of these sections are optional and others will only be processed when a specific algorithm, objective function (i.e. calibration vs. constrained optimization vs. multi-objective optimization), or feature is activated. Section 3 provides guidance on running OSTRICH in serial or parallel, and Section 4 describes various output files generated by OSTRICH. Finally, Section 5 reviews the set of examples that accompany the OSTRICH distribution and provides instructions for running them on a Windows-based machine.