Can matlab determine a constant in a formula that convert a dataset from an experiment through a trial-and-error process ?

Is it possible to write a script in matlab that determine the optimal constant in a given formula which converts a dataset from an experiment into another dataset ?

Let my try and clearify:

Im currently working on my master thesis project in biochemistry which involves enzyme kinetics. From the experiments i obtain a set of data (optical density (OD) measurements at specific time points)

Unprocessed these data will results in a curve like this:

These data needs to be process by the following formula

B(t) = (1-k)*B(t-1) + (A(t)-A(t-1)), where A(t) are unprocessed data and k is the constant in question.

The optimal value of k, will results in a dataset that looks like this

more specifically, the end part of the curve needs to be horizontal (slope = 0).

Is it possible to write a script in mathlab, that runs a loop and tries different values of k and determine the optimal value of k by comparing R-square from linear regression of the last part of the curve?

Thanks in advance.