Hopkins Research Logo
  home
  what do we do?
  who will be interested?
  how can we help you?
  what have we done?
  what is the future?
  technical information
touch the future

What do we do?



Systems of interest to scientists, engineers, manufacturers and economists have become so complex that making optimal decisions about them is now impossible without sophisticated help. Success increasingly relies upon accurate and reliable models of system behaviour – using experimental data, subject area knowledge and sophisticated empirical techniques.  The inputs to these models are those variables that the investigator has control over or can observe. The outputs are the responses that measure system performance.  Integrating such models into a collaborative design process is essential if you want the best solutions.  We will help you to create and use the best models available.

No model will capture all the subtle behaviours of a complex system, but once you have models that are 'good enough' (which can be measured if you know how), you have the ability to carry out the following with unprecedented efficiency – leading to better solutions, faster

  • Visualisation & communication of trends in system behaviour
  • ‘Intelligent’ high-dimensional, multi-response optimisation
  • Sophisticated simulation/capability studies
  • Assessment of complex trade-offs between cost & system performance
  • Measuring system sensitivities & average performance in the presence of uncertainty

Our techniques are unique in producing this valuable information in an integrated and reliable way. We can convert the intractable complexity of essentially any system into a convenient virtual surrogate that can be manipulated on a computer.  The key (and difficult challenge) is to produce models that are good enough from necessarily limited data.  For most contemporary problems this requires mathematical sophistication, original project management and problem-solving skills, experience and proprietary software tools.  Commercial software with two weeks of training are not good enough, however much some managers may hope.

 
 
"all models are wrong, but some are useful"
George Box
   
  Go to Flash site © Hopkins Research Ltd 2011  All rights reserved