具有几何约束的非等面积界面布局优化方法

目的 由于计算的难度与精度问题，在求解布局的过程中通常不考虑待布局元素面积的大小和形状的几何约束问题，导致求解的布局结果是理论上的，无法达到真正实用的目的。为提高界面布局优化方法生成布局方案的可用性，提出了一种集成启发式算法、多属性决策(MAMD)和整数线性规划(ILP)的方法。方法 首先，在获取到待布局元素间相关性、待布局元素的使用频率和面积等数据情况下，使用遗传算法(GA)通过改变待布局元素几何约束的参数生成一组备选布局方案;其次，将交互成本、GA寻优所花费的时间、有效性、效率和满意度作为评价界面布局的五个指标，根据TOPSIS、AHP和数据包络分析(DEA)等多属性决策方法对所有备选布局方案进行排序;最后，使用ILP方法获取一致性排序。结果 获得一个有效、具有高可用性的布局方案。结论 根据实验结果可知，该方法寻到的最优布局方案比原始方案有效地降低了交互成本和算法寻优的时间，提高了布局方案的可用性。

Non-equal Area Interface Layout Optimization Method with Geometric Constraints

Because of the difficulty and precision of calculation, the size of AOIs and geometric constraints are usually not considered in the process of solving the layout, which leads to the theoretical layout results and cannot achieve the real practical purpose. The work aims to propose a method integrating heuristic algorithm, multiple attribute decision making (MADM) and integer linear programming (ILP), so as to improve the availability of interface layout optimization method to generate layout scheme. At first, under the condition that the data such as the correlation between AOIs, the use frequency and the size of the area of AOIs were obtained, a set of alternative layout schemes were generated with genetic algorithm (GA) by changing that parameters of the geometric constraint of AOIs. Secondly, the interaction cost, time of GA optimization, effectiveness, efficiency and satisfaction were taken as the five criteria to evaluate the interface layout, and all the alternative layout schemes were ranked according to multiple attribute decision making such as TOPSIS, AHP and Data Envelopment Analysis (DEA). Then, the ILP method was used to obtain consistent ranking. Thus, an effective and highly available layout scheme was obtained. According to the experimental results, it can be found that compared with the original scheme, the optimal layout scheme can effectively reduce the interaction cost and algorithm optimization time, and improve the availability of the layout scheme.