近期，通信与信息工程系科技协会传来喜讯，继上次两名同学的科技论文被录用后，又有两名同学的论文被录用，分别是物联网创新工作室成员、16物联网工程专业学生杨蕊铭与王小玄同学所写的两篇论文《Modeling and Prediction of New Energy Use》与《基于平均等待时间最小化的电梯自适应调度方案设计》分别被2019 4th International Conference on Energy, Environment and Natural Resources和《创新科技》录用。
杨蕊铭，荣获过北京邮电大学优秀团支书，优秀团员等多项荣誉。在校期间积极参加校内外活动，参与多项学科竞赛并积极参加世园会志愿者活动。完成论文《Modeling and Prediction of New Energy Use》并被2019 4th International Conference on Energy, Environment and Natural Resources录用。其他学生作者：商子彦 杨凯。
论文简介如下：
This paper looks into clean energy consumption in the four states of California (CA), Arizona(AZ), New Mexico(NM) and Texas (TX) by analyzing and comparing the methods of energy consumption, the similarity and difference of their energy composition and the causes for it, and finding out the state with the optimal ways of energy consumption, and based on it, predicts the future energy composition of these states and proposes a target for interstate energy convention. And through multiple regression analysis, and the corresponding indicators of the methods of energy consumption in these states, we compare the ways of new energy consumption in these states, and analyze the difference from the perspective of industries and geographies in these states, which prepares necessary reference for the following modeling. After some basic analysis of the data, we establish a multi-attribute decision making to find a state with optimal composition of energies through the five indicators of energy composition, volume of clean energy consumption etc; and based on the analysis, we find the different characteristics of energy consumption in these states. Then we set up a GM (1,1) model to make prediction based on the data of energy consumption of the near 20 years and project energy consumption of the four states in 2025 and 2050. By means of comparing with different models, we have nearly the same conclusion: CA is a state with optimal energy combination and has best practice for future development. There in projecting the 2025 and 2050 energy consumption, we can use CA as a reference state and set such as the target for energy convention between these four states.
Key words: Multiple regression analysis, multi-attribute decision making, principal component analysis.
指导教师：李雷远、刘刚、任国芳、吴娱、张长江、 张震
王小玄，担任班级团支书职务，荣获过优秀团员、校级奖学金等多项荣誉。在校期间积极参加校内外活动，创新竞赛及世园会志愿者活动。
期刊名称：《创新科技》
指导教师：李雷远、刘刚、任国芳、吴娱
论文名称：《基于平均等待时间最小化的电梯自适应调度方案设计》
简介：在电梯群控系统中,由于存在着很多的不确定性因素,如乘客数量的未知,厅层召唤的未知性和随机性,使得电梯群控成为一个具有非线性和不确定性的复杂的多目标决策问题。本文将模糊推理和神经网络相结合来处理非线性、随机性和模糊性等问题。根据专家规则确定了进行优化调度的模糊神经网络，采用误差反向传播算法对网络进行学习。模糊神经网络融合了模糊逻辑和人工神经网络的优点,易于表达知识并且有自学习能力。将平均候梯时间、平均乘梯时间、拥挤度和能耗等作为参考评价指标，建立了多目标优化的电梯群控系统的数学模型。用Matlab对实际呼梯信号进行调度仿真，验证了算法的有效性。
关键词：模糊神经网络；电梯群控；多目标规划；模糊规则 （通信与信息工程系 李是尧供稿）
