文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Introduction The Bio-Networking Architecture Bi-weekly report #5 (Aug. 5, 2002) Adaptation
侵权投诉

Introduction The Bio-Networking Architecture Bi-weekly report #5 (Aug. 5, 2002) Adaptation

Introduction The Bio-Networking Architecture Bi-weekly report #5 (Aug. 5, 2002) Adaptation and Evolution

Introduction:

Adaptation and evolution by natural selection are among key biological concepts that the PI applies in the Bio-Networking Architecture. Analogous to a biological entity in the real world biological systems, service components (referred to as cyber-entities) in the Bio-Networking Architecture follow simple behavior rules (e.g., replication, reproduction and migration). A cyber-entity invokes its behavior based on local information (such as resource availability) and interactions with other nearby cyber-entities. These autonomous behaviors of cyber-entities in the Bio-Networking Architecture result in network applications that adapt to changing

environment.

A cyber-entity also evolves in its behaviors. When a cyber-entity replicates or reproduces with another cyber-entity, diverse behavior policies are created through mutation and crossover in their behavior policies. Since an inefficient cyber-entity exhausts energy quickly and eventually dies in the Bio-Networking Architecture, only beneficial cyber-entities are retained in the system, enabling network applications to adapt to changing environments. New Achievements:

In the area of evolution and adaptation aspects of the Bio-Networking Architecture, the PI has recently published works regarding the preliminary simulation results:

In a presentation entitled “Evolution Simulations for the Bio-Networking

Architecture” at the IEICEJ Next Generation Networks workshop [SS01], and

As a book chapter entitled, “Preliminary Simulation Study of the Bio-Networking

Architecture,” [TS02].

To investigate the potential benefits of adaptation and evolution in the Bio-Networking Architecture, the PI has developed a simulator and, through simulations, demonstrated emergent behaviors such as scalability, adaptability and survivability in a variety of network environments. Evaluation of Natural Selection and Evolutionary Mechanism

Natural selection mechanism in the Bio-Networking Architecture is designed to improve network-wide performance by removing cyber-entities that are not suited for network

environments. Natural selection is performed based on the concept of energy where inefficient cyber-entities exhaust energy and eventually die, resulting in well adapted cyber entities retained on the network. With the evolutionary mechanism that can generate diverse cyber-entities in behavioral policies with possible mutation and crossover, cyber-entities gradually increase their fitness to the environment in changing network conditions.

In the preliminary simulations, the PI has evaluated the natural selection and evolutionary mechanism in a static environment. A network is configured as a 5×5 grid topology. 8 users are assumed in the simulated network with different request generation rates ranging from 5 to 30 requests per second. One cyber-entity can process 5 requests per second and receives 10 energy units from a user in return for processing a request. Each platform node charges cyber-entities 1 energy unit for computing resources (e.g., CPU, memory, bandwidth), and also 100 energy units for platform services (e.g., replication, reproduction, migration supports). The Bio-Networking Architecture Bi-weekly report #5 (Aug. 5, 2002): Adaptation and Evolution PI: Tatsuya Suda (suda@ics.uci.edu) University of California, Irvine

第1页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email