Group Recommendation In Crowdsourced Design Based On Preference Fusion And Bilateral Matching Satisfaction

The crowdsourcing design project of complex products often requires designers to form a team and complete related tasks in continuous interaction and collaboration.Therefore,this paper mainly studies how to select design teams with similar preferences from a large number of designers,and recommends collaborative design crowdsourcing tasks to the selected design teams.Aiming at the problems of different preferences of designers,such as large differences in professional knowledge background,task experience and interested projects,a team discovery algorithm(Improvement Louvain,I＿Louvain)combining member preferences and structural similarity was proposed to consider the preferences among team members and improve the module degree index.Firstly,the similarity of preference attribute and topology structure of nodes were calculated.Combined with the nodes given by the user and their neighbors,and considering their preferences and structural similarity,the candidate nodes of the target team were extended.Then,with the candidate nodes as the core,the interest preference of the design team was mined to calculate the improved modularity,and the team division was updated and optimized.Finally,the experimental results on public data sets and crowdsourced engineering case data sets show that the module degree index of team division was improved,which verified the feasibility and practicability of the algorithm.In order to improve the accuracy of collaborative design crowdsourcing task recommendation and solve the matching problem between the crowdsourcing task and the design team,a bilateral satisfaction optimization model was established by considering the bilateral preferences of the recommendation task and the design team.Converge Multiple Strategies Sparrow Search Algorithm(CSSA)was used to solve the algorithm.The sparrow population was initialized with Tent chaotic mapping to improve the randomness of the population.The salps swarm algorithm was used to update the search formula of individual sparrow finder to enhance the global search ability and scope in the early stage of algorithm iteration.The Levy-flight strategy was introduced in the alert position,and the sine and cosine algorithm was combined to enhance the global optimization ability and the ability to jump out of the local extreme value.The results show that CSSA algorithm has better searching ability,convergence efficiency and stability,compared with sine and cosine sparrow search algorithm,traditional sparrow search algorithm,gray Wolf optimization algorithm and bat algorithm on 6benchmark test functions.And taking the charging gun of smart portable new energy vehicle as an example,the feasibility and practicability of the optimization model of bilateral preference matching satisfaction are verified.Based on the team discovery method and collaborative design crowdsourcing task recommendation method proposed above,a prototype system of collaborative design crowdsourcing task recommendation based on bilateral preference matching was designed.The prototype system realizes the reasonable construction of the design team,takes into account the task recommendation function of the bilateral preference satisfaction between the design task and the design team,and provides a more targeted,fast and convenient method for collaborative design crowdsourcing task recommendation.