A ubiquitous GIS: framework, services and algorithms, development

Abstract

In this thesis, we focus on the development of a GIS system, called Intelligent Map Agent system (IMA). We implemented the system framework, designed and implemented several system services, and designed efficient algorithms for the services. The main focus of the IMA is to provide access to services, which require the usage of spatial data and spatial information, to mobile users regardless of their location. The IMA achieves this via Services Oriented Architectures [8] which supports a distributed set of services instead of a single monolithic spatial information system. The IMA system can accommodate a large number of mobile users and services, which may be distributed across a wide geographical area. The services, as well as users, are represented by agents using JADE/LEAP technologies. The first part of the thesis is dedicated to the IMA system including a description of its architecture design and implementation. In the second part, we present algorithmic solutions to two services, intended for the IMA system. These are: (i) meeting scheduling services and (ii) shortest path services. The services are using data having temporal and geometrical constraints. We present an algorithmic solution to one variation of the meeting scheduler service, where each participant (traveling in Euclidean space) is associated with a meeting schedule, and the problem is to find a location and a time where all the participants can meet for the longest possible duration. For this problem, we present two algorithms, a geometric algorithm has an O(n log n) execution time (participants travel at the same speed) and a linear programming algorithm with an O(n) expected time (participants may travel at different speeds). When no new meeting can be scheduled among all the participants, we give an O(n4) algorithm to find the maximum number of participants. We also consider the case that all participants travel in a road network (graph). We also present an algorithmic solution to one variation of the shortest path service, which is to compute a shortest path among a set of growing disc obstacles. For the case where the disc radii grow at the same constant speed, our algorithm runs in O(n2 log n) time, which improves upon an O( n3 log n) time solution. When the discs grow at different speeds, we are able to compute a shortest path in O(n3 log n) time.