How to combine two algorithms in one by using optimization results as input for second optimization

I have Parameter estimation algorithm . This parameters are either constant or variables.

The case when these parameters are constant (under certain conditions) , it is possible to do the parameter estimation and took the results as inputs for Second optimization.

But actually these parameters generally are variables (under certain conditions)

So when I optimise First optimization , the condition my changed and passe new information

I think I now i want to modify the workflow so that:

• I optimise for "first optimization"
• Grab selected outputs from the optimisation of "first optimization"
• These outputs become initial condition or inputs for "Second optimization"
• Optimise "Second optimization".

but that feels like a cumbersome solution that might be quite slow. Is anyone able to point me towards an example or guidance on how I might tackle this in a more efficient way?

Especially if we know that this is an online optimization and require a strong algorithm such as GA, ANN or metaheuristics..

Answer 1

[Subhajyoti]

Hi @YASSER,

It is my understanding that you are trying to formulate your problem into two sequential Optimizations problems.

You can use ‘Multiobjective Optimization’ for modelling your problem. It is more efficient and coherent than running separate optimizations sequentially, especially when the objectives are interrelated. The ‘Pareto front’ finds noninferior solutions — that is, solutions in which an improvement in one objective requires a degradation in another.

The following documentation has explanation and examples of minimizing or maximizing multiple objective functions subject to a set of constraints:

https://www.mathworks.com/discovery/multiobjective-optimization.html

Additionally, you can also refer to following MathWorks documentation links to know more.

• ‘Multiobjective Optimization’: https://www.mathworks.com/help/gads/what-is-multiobjective-optimization.html
• ‘Parameter Estimation’: https://www.mathworks.com/discovery/parameter-estimation.html

I hope this helps.