%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Readme file explaining the data
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% This directory contains the relevant data published to accompany the
% publication "Horsetail Matching: A Flexible Approach to Optimization Under
% Uncertainty" in Engineering Optimization

% The funding source of this work, the UK Engineering and Physical Sciences
% Research Council (EPSRC), gives Open Access requirements such that the data
% supporting the publication is made available.

% The data provided here is intended to be sufficient for recreation of the
% results in the publication. The results in the publication arose from
% performing robust optimizations on various test problems.

% This data is therfore the software used to perform these robust
% optimizations (written in python). The source code is provided along
% with scripts that run the software on algebraic test problems.

% The same methods are used for the airfoil design problem, and there is
% enough detail given in the publication such that anyone with SU2 can also
% use the provided code to recreate the results of this airfoil design problem.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Directory Layout
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% The root directory contains scripts used to perform the robust optimizations
% the src folder contains the source files required by these scripts
% the output folder contains any output and log files created by the code

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Requirements
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% The code is all written in python (v2.7.6). Additionally it uses the modules: 
% numpy (v1.12.1), scipy (v0.13.3) and matplotlib (v1.5.1)

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% User Guide
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Familiarity with the publication is assumed. 

% To compare the influence of different targets on a test problem like in 
% Figure 6, use run_target_influence.py

% To compare density matching and horsetail matching on a gaussian
% desgin space like in Figure 7, use run_gauss_contours.py

% To run density matching and horsetail matching optimizations on a test
% problem like in Figure 8, use run_HM_optimization.py. This script can also
% run an optimization of a weighted sum of mean and variance. 

% To run a multi-objective optimization of mean and variance on a test problem
% like in Figure 9 use run_MO_optimization.py

% These scripts use the class HorsetailMatchingOpt in the file HMgradopt
% to perform density matching and horsetail matching optimizations. Parameters
% such as number of samples, number of integration points, number of
% uncertainties, integration range, bounds on design space can be sepcified
% when the class is initialized.