交通系统规划与控制教学（清华大学）Transportation Systems Analysis5.pdfVIP

Transportation Systems Analysis Problem Set 1 Due date: Monday March 12, 2012 in class Attention: Besides uploading your SAS code to the course website, you are required to submit a hardcopy of your write-up (either handwritten or printed) . 1． Case Study: Developing a Binary Choice Model (60 points) In this case study, you have to estimate binary choice models using real world data. Please refer to the data description section of this problem set for more details. 2 ． Supplemental Problem (40 points) Problem 1 (20 points) In a linear city there are two stores located at points C and D as shown in the figure below. There are 500 daily trips, total, to both stores. Two hundred trips originate from the residential areas on the left of C, and 300 trips originate from the area to the right of D. (The area between C and D is not residential.) The shoppers make one trip per day to one of the stores, either store 1 or store 2 but not both. Those who live on the left of point C are uniformly distributed between points A and B. The shoppers who live on the right of point D are exponentially distributed: f (x)  exp(x), x 0 where is an unknown constant (parameter) and x is the distance from point D (to the right). The utility to individual n from choosing store i (i = 1, 2) is given by: Uin1(distancein)2ln(sizei)in where distancein is the distance of individual n from store i, sizei is the size of store i, and in is an error term, which is independently and identically distributed Gumbel (Type I Extreme Value) across all alternatives and across the population. Given distances dAB , dBC and dCD, what is the expected number of trips taken each day to store 1 and store 2? How does the value of affect the number of trips to stores 1 and