Prototyping a general solver framework of combinatorial optimization problems on a volunteer computing environment

在项目开始时，我们有以下两个研究目标：1)评估在志愿计算环境中大规模并行化处理组合优化问题的有效性，2)评估基于jdk5.0的组合优化问题通用求解器框架。结合这些目标，我们计划开发一个求解者活动的可视化工具。当我们提出这项研究时，JXTA技术已经引起了相当大的关注。然而，对于我们的研究目标来说，这项技术似乎并不那么强大。在研究了计算环境之后，我们遇到了一个软件工具Overlay Weaver，并使用Overlay Weaver在一个覆盖网络上制作了一个原型系统。由于Overlay wave有一个分布式环境的模拟器，我们可以进行大规模的计算模拟。一方面，我们没有开发基于jdk5.0的通用求解器框架。我们应该重新考虑基于java的系统是否适合我们的目标。另一方面，我们开发了一个优秀的可视化工具，它可以显示求解困难问题实例的算法的运行时行为。如今，“志愿计算”已经成为信息技术领域的一个专业术语。在不久的将来，当志愿者计算环境变得更加成熟时，我们应该重新启动类似的项目。

We had the following two research goals at the beginning of this project: 1) Evaluating the effectiveness of a large-scale parallelization for combinatorial optimization problems on a volunteer computing environment, 2) Evaluating a JDK5.0-based general solver framework for combinatorial optimization problems. Together with these goals, we had a plan to develop a visualization tool for activity of solvers.When we proposed this research, JXTA technology had attracted considerable attention. However, this technology seemed be not so powerful for our research goal. After we investigated computational environments for our research, we met a software tool Overlay Waver and made a prototype system on an overlay network using Overlay Weaver. Since Overlay Waver has an emulator of distributed environments, we could conduct large-scale computational simulations.On one hand, we did not develop a JDK5.0-based general solver framework. We should reconsider whether a Java-based system is appropriate or not for our goals. On the other hand, we developed an excellent visualization tool which displays runtime behavior of a solving algorithm for hard problem instances.Nowadays, the term "Volunteer computing" becomes a technical term in information technology. We should restart a similar project like this project in the near future, when volunteer computing environment becomes more mature one.