Optimization

Required data (see models, for details on model definition)

• For single objective optimization without constraints you will need one column, defined as 'model'.
• Single objective optimization with constraints - one column representing the model and one model column for each constraint
• Multiobjective optimization - two or more model containing columns which are chosen in the 'Objective functions' field in the Column selector tool. This tool appears every time any of the Charts is chosen. You can add columns for constraints as well.

Available optimization methods:

• Gradient search (single objective)
• Random search (single objective)
• Exhaustive search (scanning) - (single objective)
• Genetic algorithms (single objective)
• Non-dominated Sorting Genetic Algorithm 2 (NSGA2) - Multiple objectives

See also

• Scripts
• Contours and Surfaces
• Use discrete variables
• Models
• Robust design and optimization
• Working with discrete variables
• Mixtures and linearly dependant variables
• Tutorial on model definition, optimization and robust design

Back to Data Entry