Search

In this technical paper, the authors apply a cumulative genetic algorithm and a neural network to the transportation network design problem. Constraints of the problem must be addressed during the design process to determine the best feasible solution. After introducing the algorithm, it is demonstrated in examples of restrictions encountered in practice. Then, an analysis is made of alternative constraint processing methods and the optimization process is modified within the cumulative genetic algorithm. Two groups of constraints are defined for a test network to verify the applicability of the method. Results indicate that the search process yields the best feasible solutions, confirming the use of the cumulative genetic algorithm and neural network in processing constraints.

In many urban-planning, geographic analysis, and engineering studies, a classical problem is finding the shortest paths within a bounded area, or the best path that goes around or through special regions, called barriers. In this paper we present two algorithms for finding the shortest or best paths. The first one is designed to find the shortest Euclidean path in an area represented as a polygon. In this algorithm, we first select an arbitrary path between the two points, then modify this path until it is the shortest. Its computational complexity is proven to be superior to that of previous ones. The other algorithm is for finding the best path between two locations, using a rectilinear metric in the presence of penetrable barriers. To minimize both total travel time and cost, the route can either go around or through each barrier. When selecting the best path, multiple-weighted criteria are used in the evaluation. Examples are presented to illustrate how these two algorithms work.

A prototype man-computer system was developed for use in the design of urban bus transit systems. The Urban Transportation Analysis Section (UTRANS) permits design and evaluation of many alternatives within a relatively short period of time. UTRANS is currently limited to the design of transit systems that serve a many-to-one travel pattern, i.e., many trip origins having a common trip destination. The designer uses a graphics terminal to input design data for a bus transit system. A computerized behavioral model of mode split then predicts performance of the system design. A series of performance displays are provided for use in evaluation of the design. Repeated modifications of the original design permits the generation of many alternatives and eventual selection of a particular design that has best overall performance. An assessment of the utility of UTRANS in a real-world setting awaits the results of current efforts to apply the system in the design and evaluation of transit service to the University of Washington, Seattle, Wash.

As shown by the experiences gained in recent energy crises, transit agencies should prepare emergency contingency plans to deal with possible rapid and sharp increases in the demand for transit service. In this paper, contingency plans are developed for two suburban areas, Bellevue and Federal Way, in the Seattle, Washington metropolitan area. Bus transit system designs are developed for three emergency levels: 100, 75 and 50% of the work trips to be carried by buses. Altogether, 10 plans are developed. Six assume no vehicle availability constraints, and four assume only existing Metro and school buses are available. Results indicate that bus vehicle requirements for the 100 and 75% levels far exceed likely vehicle availability levels while the 50% level could be handled with currently available vehicles (Metro fleet combined with school buses). This design work was accomplished rapidly and effectively with an interactive graphic computer program called the Transit Network Optimization Program (TNOP). It would be impossible to do a similar study using conventional transit planning techniques.

The citations in this bibliography are presented by subject: airport planning and design, automobile design, citizen participation, downtown people-mover (dpm) systems planning and design, highway facilities, highway route planning, highway safety, human figure simulation, planning and design of ride sharing systems, rearload analysis and planning simulation of outer space and aircraft operations, simulation of traffic movements, traffic signal optimization, transit facilities and operations analysis and planning, transit system planning and design, transportation data display and editing, transportation education, transportation systems planning. (TRRL)

Graphics are a powerful but often costly means of communication. Computer-drawn graphics offer a new and relatively inexpensive way to assist the communication of complex technical information to both planners and non-technical people whose full potential is now beginning to be realized. This paper discusses the utility of graphics and introduces three computer-drawn graphic techniques which may be useful, both for analysis and presentation of results, in the transportation planning process. Cenvue(s) produces a three-dimensional, perspective-view map, on which virtually any type of transportation data or performance indicator can be displayed. Vap is designed to display origin-destination travel patterns in any region. Tdn transforms a physical-distance network into a time-distance network so that effects of different speeds in the transportation network can be readily seen. The cost-effectiveness of each technique is briefly discussed and some recommendations for evaluating computer graphics techniques are provided to aid the user in further assessing their utility in the transportation planning process. /Author/TRRL/

Utilisation des techniques de la conception assistee par ordinateur pour l'etude des interactions entre transports et occupation des sols. Presentation du logiciel UFOS (Urban form optimization system)

A measure of locational efficiency, the locational imbalance vector (LIVOR) is developed and applied in the analysis of nineteen hospitals which provide medical, surgical, and obstetrical services in the Cincinnati metropolitan area. With only a few exceptions, the locational efficiency of the individual hospitals presently serving the Cincinnati area is found to be quite low. Some possible policy implications of the study's measure of locational efficiency are illustrated and its relation to a larger question, that of explaining the pattern of occupancy pressure in a metropolitan area, is also briefly discussed.