当前的网络学习存在着"信息过载"、"缺少个性化"、"缺少长远规划"这些缺陷。为了在这种情况下能根据学习者的个性化特点及资源的动态性为其提供一个优化的学习过程及其支持资源，本申请项目将首先对e-learning进行基于知识点的、以过程为中心的集成化建模，并建立模型之间的动态映射；之后，根据e-learning的特点，利用有向超图的性质，在考虑资源约束的情况下进行学习过程的结构优化变换及学习序列的提取，给出相应的优化规则和定理；根据学习者的学习目标和资源环境利用遗传算法求解优化的学习序列及学习资源；根据学习者的反馈信息利用神经网络进行学习目标、学习背景和资源环境关系的调整；通过对比来分析当学习目标发生变化或者资源环境发生变化时应该如何调整学习方案。基于上述理论开发一个e-learning建模与优化原型系统。最后，对计算机领域相关课程进行建模，并选择有代表性的课程Java程序设计的学习进行验证。

With the development of computers and networks, online learning resources are becoming richer and richer. However, each learner's learning background and objectives are different. For this reason, it is important to recommend an appropriate learning process and learning resources automatically to each learner. To optimize a long-term e-learning process according to the dynamic environment, an integrated and unified e-learning model is developed. The e-learning is formally described using elements of knowledge drawn from five perspectives: the course, the learner, the task, the learning process, and the learning resource. Next, a set of formal rules and theorems are provided to set up a dynamic map between these sub-models. Rules are given for the extraction of the optimal learning sequence and its supporting resource according to the learner's learning objectives and resource background from the available learning processes, and a web-based prototype system is developed in Java. Finally, the modeling and optimization theory is validated in computer field by an experiment in which a number of learners with different learning backgrounds and different goals are divided into two groups to learn Java. The first group used the traditional method and the second group used the method described in this paper.