The scalability of distributed virtual environments (DVEs) requires an approach for filtering and disseminating relevant data to participating members. How should an algorithm be designed for coping with sporadic interest matching, hasty entity movement, and small aura sizes in DVEs?
Liu and Theodoropoulos present a space-time interest matching technique for spotting mislaid events between discrete time steps necessary in accurate simulations. They succinctly review the strong points and downsides of the persuasive zone-based, aura-based, class-based, and hybrid interest management schemes, and the viable interest matching algorithms. In the space-time interest matching technique, a swept volume confines the movement of any haphazard virtual object along any unsystematic trajectory over a time period. Swept volumes are used to bind the pathway of auras over each time phase. A resourceful divide-and-conquer technique is used to ascertain whether or not two swept volumes truly overlay at a specific time. In a simulation, the interactions of pairs of auras among twofold successive time steps are exploited via the aura-based filtering component of the space-time interest matching technique.
The authors present a space-time interest matching algorithm that supports interest matching among multiple regions in DVEs. A sorting procedure is utilized for partitioning a multidimensional puzzle into single-dimensional problems, prior to exploring the possibility of overlapped pairs of swept volumes, and then executing the pairwise interest matching algorithm. The capability to pinpoint missing events and the computational performance of space-time interest matching were evaluated in a simulation experiment. The 3D simulation of World War II dog fighting involved two teams of virtual aircraft in aerial combat, where the information update and subscription regions of the aircraft vigorously changed. The experimental results reveal that the space-time interest matching algorithm is (1) able to capture more swept volume overlaps than the existing discrete algorithms, particularly when entities travel at high speeds, and is (2) much faster than the brute-force approaches.
Although the divide-and-conquer algorithm and the swept volume computation hinder the runtime efficiency of space-time interest matching, the authors present techniques that reliably perform interest matching at discrete periods for fast-moving entities in scalable DVEs. The mathematical ideas in this paper are timely and useful for investigating complex simulation tasks in data-exhaustive exascale multiagent system simulations [1,2]. I really enjoyed reading this insightful paper.