| Summary: | 碩士 === 國立成功大學 === 土木工程學系 === 88 === One common method to obtain the Origin-Destination (OD) information of a highway system is by license plate matching. Time-and location-stamped plate numbers are copied manually, voice recorded, photoed, or vedio taped at entrances and exits of the highway. Records taken-at entrances and exits are then paired up appropriately, yielding the desired OD information as well as its spatial distribution. Regardless of the recording method, the quality of the data is often low due to human errors, high traffic volume, or climate and environmental conditions. In practice, most of the records cannot be matched appropriately, thus only a small portion of the obtained information is useful. However, we observe that while most of the data are either incomplete or inaccurate in either plate number or its accompanying time or location stamp, most of them contain some fragments of correct data.
Therefore, accepting imperfect matching of records can help one make more use out of the data. In this research we develop a model to solve for an optimum matching for a set of survey data. First a score is assigned to each possible pair of records according to the matching quality. Factors considered at this stage include the goodness of matching between the recorded plate numbers, as well as the time and location information. Then the problem is formulated as an assignment problem, and finally solved for a set of matching that maximizes the total score. Testing with randomly generated data indicate the method is superior than the traditional method.
|
|---|
