Hi

by my mean, there are few approaches.

I suggest two different types:

Neuron network approach

big training set: (input) parameters of task - (output) optimal operation sequence

it is something as:transfer speach to text by neuron networks

genetic algorithm approach

generate operation sequence (random or based on set of rules) - analyze the found sequence - determine fitness value for it

repeat it, with basic genetic algorithm operation - crossing, mutation ...

and combine it

see a similar question:

Rafik Elaakil

Abdelmalek Essaâdi University

How can combine Genetic Algorithm and Machine Learning for a combinatorial optimization problem

thank youn Jaroslav Rojíček .

Merci.