This paper investigates a Multi-objective Optimization approach where several functions need to be optimized at the same time. We introduce a new model for efficiently mining Pareto optimal patterns with constraint programming. Our model exploits condensed pattern representations to reduce the mining effort. We design a new global constraint for ensuring the closedness over a set of measures. We show how our approach can derive high-quality non redundant association rules without the use of thresholds whose added-value is studied on a case study related to the analysis of genes expression data.