Geography Reference
In-Depth Information
the algorithm until the algorithm terminates (Goldberg 1989 ). It can be seen in
this evolution process that “survival of the fittest” is implemented by providing
more chances to those better-fit individuals so that they can be carried over to the
subsequent generation.
According to Buckles and Petry ( 1992 ), genetic algorithm has good self-learning
and self-adaptation capabilities and needs neither prior knowledge nor artificial
interference in the process of searching for the optimal solutions. Its selection,
crossover and mutation operations increase the vitality of the search process
and provide more chances to find the optimal solution. Nevertheless, like most
optimization techniques, a major problem with GA is the risk of trapping into
local optimal that we may not even know it. Over the past few decades, various
methods has been proposed to design efficient and robust GA either from the aspect
of algorithm itself like diversified encoding schemes, specific crossover or mutation
operations etc. or by combining GA with other algorithms like neural network, ant
colony algorithm and so on. For more details, readers are referred to Andre et al.
( 2001 )andRoeva( 2008 ).
13.3.2
A Genetic Algorithm for Trip Purpose Detection
We consider the detection of activity types or trip purposes from GPS tracking
data as a classification problem because different trip purposes are associated with
different conditions. In order to fulfill personal, family or social needs, individuals
assume different roles at family, among friends, and at work through performing
activities of different types associated with trips of different purposes in their daily
life. There are different classifications of trip purposes or activity types (Bhat
and Koppelman 1993). A general classification includes seven categories such
as work, study, shop, social visit, recreation, home and others (Bohte and Maat
2008 ). Alternatively, one may use 'home' as the reference point and group different
purposes of trip into 'home-based' or 'non-home based' such as home-based work,
home-based education, home-based shopping, non-home-based work, etc. (Stopher
et al. 2008 ).
Trip purposes or activity types differ on a number of aspects including where
are the trip ends or activity locations, when and for how long the activities are
conducted. Thus, the land use type of trip ends or where activities are conducted
provides important information to identify trip purpose. Similarly, trip start/end
time and the duration of activities are different for different trip purposes (Stopher
et al. 2008 ;Wolfetal. 2004 ). On the other hand, individuals' socioeconomic
characteristics such as age, gender and household composition determine the types
of activities they perform (Lu and Pas 1999 ). For examples, Wang and Law ( 2007 )
find that employed persons made fewer trips and spent less time for recreation and
maintenance activities than unemployed ones. Chen and Mokhtarian ( 2006 ) report
that individuals with high income spent more time on maintenance activities than
others did. According to the investigation of Curtis and Perkins ( 2006 ), male tend
Search WWH ::




Custom Search