Given　a　finite　set　of　alternatives,　the　multi-criteria　decision　making　problem　consists　in　ranking　each　alternative　from　the　best　to　the　worst　ones.　In　this　paper,　we　are　interested　in　multi-criteria　ranking　problems　in　the　existing　method　PROMETHEE(Preference　Ranking　Organization　Method　for　Enrichment　Evaluations).This　method　requires　the　elicitation　of　preferential　parameters　in　order　to　construct　a　preference　model　which　the　decision　maker　accept　as　a　working　hypothesis　in　the　decision　aiding　study.　A　direct　elicitation　of　these　parameters　requiring　a　high　cognitive　effort　from　the　decision　maker,　proposed　an　interactive　aggregation-disaggregation　approach　that　infers　the　PROMETHEE　parameters　indirectly　from　holistic　information,　i.e.,　training　examples.　In　this　approach,　the　determination　of　PROMETHEE　parameters　that　best　restore　the　training　examples　is　formulated　through　a　linear　program.　In　this　paper,　we　consider　the　subproblem　of　the　determination　of　the　weights　only(the　thresholds　being　fixed).　The　subproblem　leads　to　solve　a　linear　program.　Numerical　experiments　were　conducted　so　as　to　check　the　behavior　of　this　disaggregation　tool.　Results　shows　that　this　tool　is　able　to　infer　weights　that　restores　in　a　stable　way　the　training　examples.　©　2010　IEEE.