Concurrent Gradients Analysis
Concurrent Gradients Method (CGM) CGM is a new patent pending approach to performing multi-objective optimization with gradient methods. Unlike previous gradient methods which can only optimize a single objective, CGM will optimize multiple objectives simultaneously by looking at the gradients of all the objectives and combines the information to find a direction of simultaneous improvements of all objectives.
CGM also explores the design space from multiple starting points The combination of simultaneous criteria improvement and multiple starting points allows CGM to avoid local optimums which have been the problems with traditional gradient methods. CGM is unique in its ability to find global Pareto Frontiers and is able to find Pareto optimal points in as few as 5 evaluations.
- Concurrent Gradients Method (CGM)
- Dimensionally Independent Response Surface Method (DIRSM)
- Pareto Navigator Method (PNM)
- Visualization Technology
- Glossary
Concurrent Gradient Method (CGM) optimizes multi-objective / multi-discipline designs in high-dimensional and highly constrained scenarios until no futher criteria improvement is possible. Once the design is optimized to within a user input threshold, Concurrent Gradient Method moves on to the next design.