Assessing representation error in point-based coverage modeling

 Accurately representing geographic space in a digital environment continues to confound and challenge researchers. Carrying out spatial analysis in a setting where geographic representation is subject to change poses problems to be addressed. In this paper we examine spatial representation in the context of coverage-based location modeling. A geographic region can be represented in a variety of ways. We present an evaluation of spatial representation in the location of facilities that provide coverage oriented services. The analysis shows that the modeling results are sensitive to how spatial demand for service is represented in a digital environment. We develop an approach for evaluating representational appropriateness. This research contributes to spatial analysis integrated in geographic information system environment. Received: 10 January 2001 / Accepted: 8 April 2002     

Map projection errors in the Weber problem

 When demand points are given as a planar map where projection method is explicitly stated, we usually know the latitudes and longitudes of these points from the map. Then we can solve the Weber problem on the globe, and we do not suffer from errors. This paper analyses how cylindrical projections cause distortion in the Weber problem when demands are distributed on the Northern Hemisphere. First, we demonstrate that planar solutions are always located south of the spherical solution if the Mercator projection, the equirectangular projections with standard parallels near the demands, or the equal-area projection with the same characteristic is chosen. Second, we verify that this geographical tendency is inclined to hold when the demand points, are distributed symmetrically, widely or toward the north. Received: 15 August 2001 / Accepted: 20 April 2002 This paper was partially written while the first author was visiting the Department of Geography at the Catholic University of Louvain, Louvain-la-Neuve, Belgium [1993–1994]. He is grateful for the hospitality of this department. An earlier version of this paper was presented in 1994 at the Seventh Meeting of the European Operational Research Working Group on Locational Analysis in Brussels, and in 1996 at the Fifth World Congress of the Regional Science Association International in Tokyo. The authors would also like to thank the participants as well as three anonymous referees for their constructive comments.     

1998年副热带高压中期数值预报产品的误差分析

通过统计和动力学分析结果的验证,可以得到:(1)欧洲中期数值预报产品是预报副高中期活动的重要依据,并且预报值的平移订正不可能对预报准确率有多大的改进。(2)不考虑天文潮汐可能是1998年欧洲中期数值预报产生较大误差的重要原因。(3)进一步分析找到二条判断欧洲中期数值预报因不考虑天文潮汐将发生较大误差的指标: ① 当交点月奇异点月下点 54.7 线与当日20时500 hPa的 50～150 E 范围内副热带急流槽线相重合,并与等温线夹角 α>60 °时,那末若以奇异点后0～2天作为初日,欧洲中期数值预报120 E 588线北界预报将出现连续三天大于等于2个纬距的连续性错误(7/9),反之,将不出现连续性错误(13/13);② 在不满足上述情况下,如果54.7 跋咄ü至迅备咧涞脱骨保?则在交点月奇异点当日出现离散性错误(5/7),反之,将不出现错误(6/6)。(4)1999～2000年汛期实际应用结果基本上证实了上述二个指标。     

GPS测量基线选取对平差结果的影响

基线是GPS测量网解算中的基本量，解算过程中占有重要位置。本文介绍了GPS测量中基线的选取方法，研究了GPS网中独立基线的选取对解算结果的影响。     

自适应BP算法及其在河道洪水预报上的应用

提出一种改进的BP算法,即自适应BP算法。该方法采用两种策略:一是在权重修改公式中加动量项;二是学习率随总误差的变化作自适应调整,亦即总误差增加时,学习率将减小,反之学习率增大。以上两种策略能有效的抑制网络陷于局部极小并缩短了学习时间。实例研究表明,该算法用于河道洪水的预报,能取得令人满意的结果。     

长度计量仪器测距指标处理系统的开发运用

测绘行业近几年发生了巨大的变化 ,传统的测量方法大多被数字化手段所代替 ,然而部分测绘仪检部门在处理测距仪器检定数据时有的采用自行编制的功能不甚完备的程序进行计算。针对这种情况 ,作者开发了长度计量仪器测距指标处理系统 (KRA) ,该系统能较为全面地满足光电测量仪器检测数据的后期处理。本文就该系统的开发和运用作了详细阐述。     

坐标参数加权平差法在线路补测中的应用

在线路施工的补充测量控制中,对先期控制点老坐标的如何使用,关系到新坐标的网形精度和与老坐标的差异程度.介绍了对先期控制点坐标进行赋权参加平差的方法,并给出评定结果的指标.用实例分析验证了这种方法的有效性.     

二维随机点的熵误差指标

根据信息论基本原理，提出了二维随机点的熵误差指标，该指标与以往的误差指标不同，它是一个在熵意义下唯一确定的，与置信概率无关的客观指标，把它作为二维点位置不确定性的度量指标具有特殊优越性。