Article Preview
TopSchool Bus Routing Problem
The School Bus Routing Problem (SBRP) is the objective of a fleet of school buses to efficiently pick up and drop off students to and from school (Li & Fu, 2002). These objectives must abide by specific constraints, such as bus capacity, the number of available buses, school start time, and maximum allowable travel time (Leiva, Muñoz, Giesen, & Larrain, 2010; Spasovic et al., 2001). Several studies have shown that it is almost impossible to efficiently design bus routes that account for all of these desired factors (See Li & Fu, 2002; Spasovic, Chien, Kelnhofer-Feeley, Wang, & Hu, 2001). With the multitude of the SBRP’s variables, it is challenging to balance the objective of the school system in reducing costs and the objective in maintaining an acceptable level of equity for the bus riders (Spasovic et al., 2001).
In some specific case studies, SBRP is also related to other parameters, including the Multiple Traveling Salesman Problem, Vehicle Routing Problem, and urban bus routing (Li & Fu, 2002; Park & Kim, 2010; Schittekat et al., 2013). While the Multiple Traveling Salesman Problem is a classic case for a mathematical solution using Operations Research techniques (Spasovic et al., 2001), the SBRP is not easily optimized via mathematical approaches due to the complexity of the problem, the number of constraints, and sometimes conflicting variables to optimize (Li & Fu, 2002; Spasovic et al., 2001). As a result, heuristic approaches are often used to supplement purely automated algorithms (Angel, Caudle, Noonan, & Whinston, 1972). Many school systems even neglect mathematical and GIS solutions completely. For instance, school systems in Hong Kong simply use transportation official’s intuition to manually design each route, as well as place each bus stop (Li & Fu, 2002).
The SBRP is made up of a number of different parameters. While some researchers have taken holistic approaches to address all of the aspects at once (Desrosiers, Sauvé, & Soumis, 1988), others have attempted to address them somewhat independently (Newton & Thomas, 1969; Park and Kim, 2010), even though they are highly related. Treating the parameters with some degree of independence reduces the complexity and allows for specific local constraints and variables to be factored in (Park & Kim, 2010).