This paper describes a heuristic combination based genetic algorithm, (GA), for tackling dynamic job-shop scheduling problems. Our approach is novel in that the genome encodes a choice of algorithm to be used to produce a set of schedulable operations, alongside a choice of heuristic which is used to choose an operation from the resulting set. We test the approach on 12 instances of dynamic problems, using 4 different objectives to judge schedule quality. We find that our approach outperforms other heuristic combination methods, and also performs well compared to the most recently published results on a number of benchmark problems.