The optimization suite of Isight comprises several major class of optimization algorithms classified into gradient-based, exploratory and direct search techniques. Each technique has its own set of algorithms that suits and serves different applications. Design space for each algorithm is constructed out of the input variables or design parameters. These values are varied based on a factor for each algorithm and then the feasibility of the same are checked against the penalty and objective function.

The design value population of an optimization algorithm depends on the initial value, lower & upper bound values and constraints. On top of all, there is a factor in each algorithm which determines the design value for the successive iterations. Most of the time when we choose to work with a particular optimization algorithm one of the questions that generally surface would be how the design space would be constructed, what would be the values of chosen design input variables, how and on what basis these values are populated?

With the information available on these variation factors for each algorithm, little could be inferred on the possible design input values. It is an attempt here to explore in detail how this factor works for different optimization algorithms to populate the input design values and to understand the pattern with which each new design value would be populated.