The strong convergence of a KM–CQ-like algorithm for a split feasibility problem