Multi-attribute taxi logistics optimization

According to U.S. government surveys, 12% of Americans used taxi service in the previous month' and spent about $3.7 billion a year for cab fare.2 Taxi service is one of the major modes of public transportation. Despite providing services 24 hours a day, driving relentlessly with an empty taxicab in search of passengers and answering dispatch calls instantaneously, taxi service is ranked the most unsatisfactory mode of transportation by the public. Charging higher fares than other major modes of transportation and averaging 10 to 12 hours work day, taxi drivers have a difficult time to earn a sustainable income. Approximately half of all the taxi mileage is paid mileage; this means a significant portion of a taxi's time and fuel is spent on non-revenue generating activities, i.e. without passengers. Current taxi allocation is inefficient. The number of taxis and the geographical service areas which they serve are heavily regulated in most cities. With limited competition and strict regulations, taxi service suffers with customers having to endure long wait times and inferior services. The current taxi systems in most U.S. cities may be greatly improved from their current state. This thesis investigates the factors of inefficiency in the current taxi system, reviews previous taxi efficiency studies, and suggests possible solutions. After extensive literature reviews and field research, a computer simulation model has been built in the MATLAB environment. This computer model tests various attributes that affect logistic optimizations for taxi services. In particular, the effect of taxi fleet size, the quantity of hotspots, and the concentrations of customers at hotspots are analyzed in detail using the ' Bureau of Transportation Statistics. October, 2003. http://www.bts.gov/programs/oinnibus surveys/household survey/2003/october/ 2 Schechner, S., Cranky Consumer: Hiring a Taxi During Rush Hour, The Wall Street Journal, April 26, 2005.