The strong branch option uses a more intensive branching strategy during the first n levels of the branch-and-bound tree, where n is the value of the STRONG parameter. During these initial levels, LINGO picks a subset of the fractional variables as branching candidates. LINGO then performs a tentative branch on each of the variables in the subset, selecting as the final candidate the variable that offers the greatest improvement in the bound on the objective. Although strong branching is useful in tightening the bound quickly, it does take additional computation time. So, you may want to try different settings to determine what works best for your model.

The default setting is 10 levels.