Abstract:    A R&D team composed of members with different skills and knowledge fields can take advantage of the cross-disciplinary expert′s knowledge and intelligence to handle the contradiction between individual knowledge and overall tasks, thus enabling them to understand the R&D tasks more comprehensively. However, while the knowledge heterogeneity of team members benefits from multi thinking and perspectives, it is often difficult for the team to quickly reach consensus on the goals and completion methods of R&D task, which leads to task conflicts and even prolongs the period of product development. Therefore, exploring and studying the important factors affecting the efficiency of task consensus in knowledge heterogeneous teams has gradually become the focus in academic circles. The consensus of team members on task-related content is a step-by-step cognitive process of mutual communication between team members. For knowledge heterogeneous teams, the essence of consensus formation can be viewed as learning-based knowledge coordination activities among team members. 
    In order to reveal the specific process of coordination activities on the task consensus efficiency of knowledge heterogeneity team from micro perspective, the paper discusses the realization premise of knowledge coordination based on the construction of team learning mechanism, including the heterogeneous characteristics of team members′ knowledge and the characteristics of members′ roles; then starting from the selection of knowledge learning objects and knowledge learning of team members, the specific process of knowledge learning of team members and its possible impacts on team members′ task consensus are studied. Furthermore, based on the research results of group decision-making, opinion dynamics and the Deffuant-bounded confidence hypothesis, a task consensus model under the influence of members′ knowledge level was constructed to reveal the micro-effect mechanism of team knowledge coordination on task consensus efficiency. The task consensus model under the influence of the changes in the members′ knowledge level is not only an extension of the classic bounded confidence model, but also the members′ understanding of the task that includes learning knowledge, perfecting opinions, suggesting ideas, making choices and reaching consensus, which is consistent with the five-factor structure process proposed by Andres and Shipps. Therefore, the construction of this model is advanced and realistic.
    In order to perform simulation analysis on the task consensus model constructed above, firstly we randomly generate corresponding values according to the model constraints to construct a matrix of opinion values, knowledge levels, learning capabilities, and communication capabilities; After entering the knowledge learning process, each team member selects the most appropriate member from his or her neighborhood as the learning object according to the learning object selection conditions and knowledge learning strategies; Finally, each member randomly selects a certain knowledge dimension to learn and subsequently his understanding of tasks will gradually deepen as the knowledge levels are changed. After the calculation of quite a few sets of simulation data, it was found that, on the whole, a relatively stable evolution regularity has been shown. Using the simulation method, the influences of member learning ability, trust threshold level, knowledge learning strategy and learning object searching radius on the task consensus efficiency of knowledge heterogeneity team are studied. 
    Our findings from the task consensus model based on knowledge coordination show: (1) The higher learning ability of team members enables the team to quickly reach a high degree of consensus in a short period of time, which has a positive impact on the team task consensus efficiency; and with the gradual evolution of time, the degree of this positive impact becomes increasingly apparent. (2) The trust level is positively related to the team task consensus efficiency. However, when the trust level exceeds a certain threshold, its influence on the consensus efficiency of team tasks gradually becomes more moderate. (3) Different knowledge learning strategies adopted by members will have significant differences in the research of team task consensus efficiency. Specifically, the use of hybrid strategies can most effectively improve the team task consensus efficiency, and it can have a more decisive effect on the team′s task consensus efficiency in the middle and late stages, followed by practical strategies and utilitarian strategies. (4) The learning object searching radius and neighborhood type also have a significant impact on the efficiency of team task consensus. The study found that the efficiency of team task consensus in the neighborhood type of the second-order Moore, second-order Von Neumann, first-order Moore, and first-order Von Neumann decreases in order.
     In view of the above results, this paper recommends that in addition to short-term positions and skills training, enterprises should also establish long-term mechanisms, such as building a learning organization, and creating and maintaining the learning atmosphere of independent learning, lifelong learning, active exploration and independent thinking within the enterprise. At the same time, enterprises have to continuously cultivate the meta-cognitive ability of team members to stimulate their higher-level thinking forms, so as to effectively improve employees′ knowledge reserve and learning ability level in daily work practices, promote members′ absorption of new knowledge and better promote the team to complete R&D tasks. Considering the importance of members′ learning ability, we recommend that companies should pay great attention and take effective measures, such as assigning challenging work tasks to members on the basis of clear authorization goals, allowing members to internalize what they have learned and learn to think systematically in the process of completing tasks, so as to continuously improve their learning ability. To a certain extent, this research reveals the regularity of task consensus in knowledge heterogeneity team based on knowledge coordination. On the basis of expanding the classical coordination theory, not only will the paper reveal the key aspects affecting task consensus efficiency more concretely and deeply, but provide targeted theoretical and methodological guidance for relevant management practices of enterprises.

Key words:  knowledge heterogeneity team, task consensus efficiency, knowledge coordination, team learning