In this study, a genetic algorithm has been applied to a multi-objective shape design optimization problem concerning a heat exchanger configuration close to practical applications. The characteristic Pareto front associated with this problem has been obtained within a very reasonable computational time. This model problem is based on a description of the blade geometry using only four parameters. This is clearly a very simple description that could be refined. Computing times can be further reduced by using parallelization on up to 15 nodes if needed, as demonstrated in this paper. More complex, practical industrial cases are solvable using on one side appropriate modeling and simplification of the problem and on the other side parallelization.