多目标优化视角下在线学习群体形成方法

Online Learning Group Formation Method From the Perspective of Multi-objective Optimization

形成既能满足教师教学实施需求,又能得到学习者认可的在线学习群体是影响在线协作学习效率的重要因素.多目标粒子群算法和遗传算法应用于在线学习群体形成领域是目前的研究热点.然而,利用多目标粒子群算法解决在线学习群体形成问题时存在多样性差,容易陷入局部最优等问题;运用遗传算法解决在线学习群体形成问题时,则需要以耗费大量时间为代价.针对以上问题,提出了多目标优化视角下在线学习群体形成方法:首先根据学习者的多维个性特征建立在线学习群体形成MOLGFM模型(Multi-objective Online Learning Group Formation Model),其次针对形成模型的多目标优化特征,将多目标粒子群算法和遗传算法相结合提出了GAMOPSO(Genetic Multi-objective Particle Swarm Optimization)算法,最后采用GAMOPSO算法求解MOLGFM模型,提出多目标优化视角下的在线学习群体形成方法GAMOPSO-FA(Genetic Multi-objective Particle Swarm Optimization-Formation Approach).实验表明,相比采用经典算法的在线学习群体形成方法,所提GAMOPSO-FA方法形成的在线学习群体符合度更高,形成速度更快.

The formation of an online learning group that can not only meet the teaching implementation needs of teachers,but also be recognized by learners is an important factor affecting the efficiency of online collaborative learning.The application of multi-objective particle swarm algorithm and genetic algorithm to the field of online learning group formation is the current research hotspot.However,when using the multi-objective particle swarm algorithm to solve the problem of online learning group formation,there are problems such as poor diversity and easy to fall into local optimization;when using genetic algorithm to solve the problem of online learning group formation,it needs to spend a lot of time at the cost.In response to the above problems,an online learning group formation method from the perspective of multi-objective optimization is proposed:First,establish a MOLGFM(Multi-objective Online Learning Group Formation Model)model based on the multi-dimensional personality characteristics of learners,and secondly aim at the multi-objective formation model.To optimize the characteristics,the GAMOPSO(Genetic Multi-objective Particle Swarm Optimization)algorithm is proposed by combining the multi-objective particle swarm algorithm and genetic algorithm.Finally,the GAMOPSO algorithm is used to solve the MOLGFM model,and the online learning group formation method GAMOPSO-from the perspective of multi-objective optimization is proposed.FA(Genetic Multi-objective Particle Swarm Optimization-Formation Approach).Experiments show that compared with the online learning group formation method using classic algorithms,the online learning group formed by the proposed GAMOPSO-FA method has a higher degree of conformity and a faster formation speed.