Introduction

Crowdsourcing is emerging as a powerful paradigm in human problem solving techniques to the Open World. It helps to solve a wide range of tedious and complex problems in various enterprise applications. It allows enterprise to use the creativity and the imagination of Internet users to provide a solution for an issue. We introduce a type of crowdsourcing called Socialcrowd based on the efficiency of social network to outsource a task to be performed by people on demand on the social/professional network instead of an open world as Mechanical Turk is doing. As related work one can cite [3] which studies the max/top-k and grouping database problems in the crowd sourcing setting. where the criteria used for grouping and ordering are difficult to evaluate by machines but much easier to evaluate by the crowd. In [1], we developed an approach discovering hidden relationships between crowd members in social network and building only one team that may solve the outsourced problem, in [2], we extended our previous work in order to discover a set of collaborative et competitive team compositions answering the business call. In this paper we propose to discuss a data leak free team composition method which is based on heuristics