About us    

Services

References

News

Contacts and CVs





About us

Key markets
Pulp and Paper
Panelboard
Energy

Automation and Process control

Products
Controller Series
Predictor Series
Multisimplex 
 -Pricing information
Unscrambler

Download

LubeRight

Key application fields
Process Optimization
Environment
Laboratories

Support
FAQ

Ordering

Distributors
Partners

Links

Search

Sponsor children

 

Can MultiSimplex solve any optimization problem?

MultiSimplex is a versatile tool that can solve many optimization problems, but there are also certain limitations to be aware of.

MultiSimplex assumes that the control and response variables are continuous variables. Discrete variables, that follow a meaningful scale with several steps, can often be treated as continuous. True categorical variables are however difficult to include in an optimization study with MultiSimplex and user interaction is required in every step.

Excessive noise is another obstacle for a successful optimization project. Often this can be overcome with an increase in step size and/or repeated trials.

Other frequently asked questions:

  1. I want to produce "good quality", but how do I specify the optimization objectives?
  2. How can I optimize, when the specifications do not allow any change in conditions?
  3. Can the system be too complicated for MultiSimplex?
  4. The process is not stable, how is it possible to optimize?
  5. Is MultiSimplex necessary if there is only one important variable affecting the response?
  6. Are our skill and know-how of any value?
  7. Is there any industrial experience with these methods?
  8. How does the MultiSimplex methods compare with ordinary statistical methods?

Back to Optimization Methods Introduction

Send mail to info@grabitech.se with questions or comments about this web site. Legal stuff. 
 



© Grabitech Solutions AB