The integration of multiple criteria decision making (MCDM) and geographic information system (GIS) for transit planning in Honolulu

Alternative methods for determining rail transit alignments and locations of stations are developed. The models utilize a uniform 0.1 square mile grid map system and multicriteria method, the Analytic Hierarchy Process (AHP). Although the study area is located in Honolulu, Hawaii, the approach is intended to facilitate decision-making process involving alignments or site selection in general. The vector based grid system is developed to overcome thematic maps irregular shapes and sizes problems. The grid size is within walkable distance and provides detailed demographic, social, and environmental information. The first method, the Data Driven Model, is a modernistic approach utilizing data from the grid system and expert's preferences. Locations of possible alignments are analyzed using the Directional Decision Choice Model (DDCM). The optimum alignment is selected using the AHP benefit/costs ratio or a direct selection approach. Locations of stations at grid level are identified on the optimum alignment by calculating the AHP index of their unique benefit criteria. The second method, the Community Model, attempts to increase public influence in deciding locations of rail transit alignment and stations. A set of criteria are used reflecting the Federal Transit Administration New Starts Program guidelines and other literature with regards to the optimum locations of transit alignments and stations. These criteria are structured in a unified framework following AHP principles. Community members are asked to express their preferences on these criteria by answering one of three different questionnaire types: paper, spreadsheet, or web based instrument. Responses by different respondents groups are used to generate composite AHP based indexes. This method is tested on the alternative transit alignments proposed by the City and County of Honolulu to identify the optimum alignments and station locations. Outcomes from both models are compared with the proposed transit alignments and locations of stations to see the level of agreements among those methods based on various perspectives. The study concludes with a discussion of the result, study limitations, and directions for further research.