In　current　days,　the　business　environment　is　more　competitive　and　changing　faster　than　ever　before,　so　it　is　important　for　employees　to　learn　new　knowledge　related　to　their　business　quickly　and　efficiently.　Meanwhile,　the　personalized　e-learning　system　makes　it　possible　for　employees　to　discover　a　personalized　learning　path　in　order　to　improve　the　learning　efficiency.　For　business　in　a　company,　it　is　often　required　a　group　of　people　to　complete　a　particular　business　task,　however,　the　group　members　are　not　equipped　with　the　same　level　of　prior　knowledge　related　to　the　task.　Therefore,　discovering　a　groupized　learning　path　is　important　and　necessary　for　employees　working　under　collaborative　business　environment.　In　this　paper,　we　focus　on　how　to　trace　the　roles　of　one　employee　in　a　group　in　order　to　establish　the　relationship　between　it　and　his　knowledge　level,　and　addressing　the　problem　of　improving　the　learning　efficiency　for　a　group　of　employees　rather　than　a　single　person　under　the　collaborative　business　environment.　Based　on　the　proposed　model,　an　RLPD　(role-based　learning　path　discovery)　algorithm　is　proposed　to　optimize　the　process　for　a　group　of　employees　to　grasp　new　knowledge　related　to　a　certain　business　task　more　efficiently　and　effectively,　and　its　superiority　has　been　proved　by　our　experimental　works.　©　2011　IEEE.