Flexible Multipath Search Algorithm for Multipurpose Location-Based Activities

Abstract

information itself without other types of value-added information, such as price and availability data. Imagine that one needs to purchase a dozen bottles of water and wants to find a shop via LBS. Mobile Yellow Pages might be able to provide information about the nearest shop (A) from one's current position. Car navigation services could describe the shortest path to get there, but no existing LBS model can tell the consumer which is the smarter option if there is another shop (B) that sells the same bottled water for a dollar less than A and is located only 3 min farther away.Location-related technologies have already advanced enough to be implemented for this type of service with high accuracy. Real-time or near real-time traffic and transit information systems, including parking information, are becoming available through various ITS projects conducted on a metropolitan scale. Numerous point-of-interest (POI) databases have already been constructed by private web service companies; for instance, location information services of POIs are currently provided by Google Local. Price information can be accessed through the many web pages of large brick and mortar retailers as well as Internet shopping sites. Aggregated pricing services also exist. For example, Gasbuddy.com allows users to compare near real-time (time-delayed by just a few hours) gas prices for 174 local areas in the United States.The largest hurdle in deploying user-friendly services, such as concierge services, on a widespread commercial scale is the lack of efficient solution-search algorithms. Because the problem is a complex form of the traveling salesman problem (TSP), which is an NPhard problem (3, 4), the success of this service depends on how efficiently and accurately the primary algorithm provides a response to a user's query. The problem definition and the solution algorithm are discussed next. EXISTING SEARCH ALGORITHMS: BRIEF REVIEWMany solution algorithms exist for the problem of determining traveling routes given some constraint criteria, including those by Potts and Oliver (5), Yen (6), Shier (7), Sheffi (8), Kaufman and Smith (9), and Rillet and Park (10), to name a few. All of these algorithms, however, were developed for solving for the best or optimum transportation routes without consideration of the possibility of multiple-purpose travel intended for multiple activities.Algorithms developed by Yen (6) and Shier (7) are not specifically aimed at solving multiactivity models; however, they do search for multiple possible paths to a specified destination. The algorithms are similar but differ in that Yen's (6) algorithm can suggest paths regardless of whether activities A and B are to be engaged in together. Kim (11) developed a multiactivity LBS model. Kang et al. (12) evaluated genetic algorithms (GAs) as a possible means to solve the An algorithm is developed that provides not only least-cost alternate routes but also ones that are close to the least-cost routes for users to select from multiple paths that route through point...

A dissimilar alternative paths-search algorithm for navigation services: A heuristic approach

“…They should, therefore, continue to select the least shared length path among those that satisfied the allowable travel cost as the next alternative path. Jeong et al (2007) developed a heuristic algorithm to select K paths including a specific activity out of the candidate set of alternatives to consider users' various needs. Jeong et al (2007) were able to establish a candidate set quickly by executing the shortest-path algorithm only once.…”

“…Jeong et al (2007) developed a heuristic algorithm to select K paths including a specific activity out of the candidate set of alternatives to consider users' various needs. Jeong et al (2007) were able to establish a candidate set quickly by executing the shortest-path algorithm only once. As in Park et al (2002), they, however, did not consider the duplication rates among the alternative paths.…”

“…The search algorithm for navigation services should be able to provide alternatives that satisfy users' needs efficiently. Therefore, the algorithm developed in this paper selects the alternative path out of candidate alternatives based on Jeong et al (2007) and includes user-specified allowable travel cost and duplication ratios.…”

“…On the other hand, drivers may prefer some other attributes, such as familiarity, driving convenience, and fewer right/left turns, over saving just one minute of driving time. The necessity of providing the multiple paths has been also presented by some literature such as Lee (1994), Park et al (1997), Lim and Kim (2005) and Jeong et al (2007).…”

“…Most existing navigation services, however, do not provide multiple paths so that users can choose the path that best fits their needs. Instead, existing services provide the user with a path, referred to as the 'optimal path,' which is based on a proprietary, hidden algorithm (Park et al, 2002;Mohamed and Abdalla, 2004;Jeong et al, 2007). For example, existing navigation services provide only an 'optimal' path, even though the travel time difference between the 'optimal' path and the 'alternative' path is only one minute.…”