A Machine Learning-Based Approximation of Strong Branching

A new generic approach to variable branching in branch and bound for mixed-integer linear problems by imitating the decisions taken by a good branching strategy, namely strong branching, with a fast approximation.