User-Centric Multiobjective Approach to Privacy Preservation and Energy Cost Minimization in Smart Home
User-Centric Multiobjective Approach to Privacy Preservation and Energy Cost Minimization in Smart Home
复制标题
DOI：
10.1109/jsyst.2018.2876345
复制
发表时间：
2018-11
影响因子：
4.4
通讯作者：
Hsuan-Hao Chang;Wei-Yu Chiu;Hongjian Sun;Chia-Ming Chen
中科院分区：
文献类型：
--
作者：
Hsuan-Hao Chang;Wei-Yu Chiu;Hongjian Sun;Chia-Ming Chen
考虑到住宅隐私与能源成本之间的权衡，本文研究了智能家庭能源管理。提出了一种将能源成本降至最低并最大化隐私保护的多主体方法。该方法导致了一个多目标优化问题，其中两个目标在单独的维度中被解决。相应地，开发了一种使用随机搜索的混合算法，该算法采用随机搜索来进行家用电器的功率调度并使用确定性电池控制。所提出的方法可以避免传统加权和方法的一些缺点，以进行多目标优化：在不同单位中目标的组合，加权系数的启发式分配以及对用户偏好的可能错误陈述。与现有的有关住宅用户隐私的研究相反，该研究假设设备可促进算法开发的可控性有限，此方法涉及在智能家居中使用柔性设备的使用。模拟表明，所提出的方法可以维持合理的能源成本，同时在明智的水平上稳健地保留用户隐私。它的收敛速率可与现有的多主体进化算法相媲美，而所提出的方法可以提高收敛水平。提出的方法可扩展到一组智能房屋，达到了较高的峰值与平均比率，这对基础电网的稳定性有益。
This paper investigates smart home energy management in consideration of tradeoffs between residential privacy and energy costs. A multiobjective approach that minimizes energy costs and maximizes privacy protection is proposed. The approach leads to a multiobjective optimization problem in which the two objectives are addressed in separate dimensions. A hybrid algorithm that employs a stochastic search for power scheduling of home appliances and uses deterministic battery control is developed accordingly. The proposed approach can avoid some drawbacks faced by conventional weighted-sum methods for multiobjective optimization: the combination of objectives in different units, heuristic assignment of weighting coefficients, and possible misrepresentation of user preference. In contrast with existing studies on residential user privacy that assume limited controllability of appliances to facilitate algorithm development, this approach addresses the use of flexible appliances in smart homes. Simulations reveal that the proposed approach can maintain a reasonable energy cost while robustly preserving user privacy at a sensible level; its convergence rate is comparable to existing multiobjective evolutionary algorithms while the proposed approach yields a better level of convergence; the proposed approach is scalable to a group of smart houses, achieving a superior peak-to-average ratio that is beneficial to the stability of the underlying power grid.