基于收益机制发布/订阅系统时间约束保障技术

发布／订阅系统技术具有异步、松散耦合和多对多通信的特点，有着广阔的应用前景．但是，已有的发布／订阅系统技术不能满足动态环境下有延迟需求的应用要求．针对时间约束问题，扩展了发布／订阅系统的语法，建立了延迟模型，提出了一种基于收益机制的分布式发布／订阅系统时间约束保障技术和使系统获益最大化的调度算法MTEP（maximumto tal earning priority），其特点是能够满足订阅者和发布者指定延迟约束的需求，通过与订阅者商定的价格和违约成本信息来有效地利用网络带宽，适应网络环境的动态变化．实验结果表明，该调度策略和FCFS（first come first service）、最短时间优先和固定优先级等传统策略相比，可使订阅者接收到的有效事件明显增多，并使系统收益显著改善．

a constraint solver and its application to path feasibility analysis

Many testing methods are based on program paths. A well-known problem with them is that some paths are infeasible. To decide the feasibility of paths, we may solve a set of constraints. In this paper, we describe constraint-based tools that can be used for this purpose. They accept constraints expressed in a natural form, which may involve variables of different types such as integers, Booleans, reals and fixed-size arrays. The constraint solver is an extension of a Boolean satisfiability checker and it makes use of a linear programming package. The solving algorithm is described, and examples are given to illustrate the use of the tools. For many paths in the testing literature, their feasibility can be decided in a reasonable amount of time.