摘要
在动漫制作的过程中,渲染往往耗费大量的时间和资源,因此已成为了影响动漫制作效率的重要因素。分布式动漫渲染系统是一个动漫渲染服务平台,它能够实施高效大规模的渲染。为了尽可能地扩大系统中的动态资源,提出了集中式的激励机制和基于代理的激励机制。通过随机Petri网建模仿真方法对两种机制的性能进行分析对比,得出了"随着离线处理速率的增加、基于代理的激励机制优于集中式的激励机制"的结论,为提高渲染系统效率而选择合适的激励机制提供了量化依据。
In the process of animation production,rendering often consumes a lot of time and resources.Therefore it has become an important factor to affect the efficiency of animation production.Distributed animation rendering system is a platform to provide animation rendering services,which can implement large-scale rendering with high efficiency.In this paper,in order to expand the dynamic resource of system as much as possible,a centralized incentive mechanism and an agent-based incentive mechanism are proposed respectively.Furthermore,the analysis and comparison of the performance of two proposed mechanisms are carried out by SPN(Stochastic Petri Nets) models and numerical simulations.As a result,the conclusion that with the addition of disposing offline's firing rate,agent-based incentive mechanism is better than the centralized one is derived thereby.It offers theoretical foundation for choosing proper incentive mechanism to improve the efficiency of animation rendering systems.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2008年第S2期161-166,共6页
Journal of System Simulation
基金
国家863计划项目(2008AA01A318)
北京市影视
动漫等文化创意产业研发与人才培训基地子课题(HG0837)
关键词
分布式
动漫渲染
激励机制
随机PETRI网
distributed
animation rendering
incentive mechanism
Stochastic Petri Nets
