Hey,

for control problems gradient information with respect to an objective function can be helpful. An example of an objective function: the deviation of the actual state (pollution) to an desired state. This is often called "inverse design". To reach this gradient (or better information about what to do to reach the desired state)  in a deterministic way, usually two ways exist: the direct or the adjoint approach. Whereas the direct approach scales with the number of degrees of freedom (I don't know your size) the adjoint ansatz scales with the number of objective functions (should be one). Therefor, I think adjoint analysis could be meaningful.

Sincerely

Dear Professor Niklas Kühl,

Thanks for your suggestions. The problem starts when we can not achieve the desired state. Suppose we can survive 150 to 200 PM2.5 level. However the pollution goes beyond  this level means 300. In this case how to control this immediate effect of environment. Any possible solution is there or any tool which can predict that one can simulate this effect maximally 50 deviation or more. Some alternative solution exists or any how we need to focus on planting...

You can try uncertainty measurement with Fuzzy Logic and Fuzzy soft sets

for potential parameters.

What is the purpose of optimizing potential parameters?

Dear Prof. Kalyani,

Purpose is to find a seized boundary or interval for those parameters and their reduced effect on air pollution. I want that how much air pollution can be controlled based on their potential parameters using any mathematical model. 

Potential parameters always vary with the cities.I have worked on this, you can contact me.Moreover, let us think about developing a product out of it.

Feel free to talk to me.

Dr. DADDA The BONG