|dc.description.abstract||Processor speed has been growing at an exponential rate over the past 50 years. Computers are getting smaller, cheaper and faster. Over the past 30 years, with the growth of the internet, new forms of decentralized distributed computing architectures have emerged. The emergence of distributed architectures has led to the creations of distributed computing systems and a new field of research.
Distributed computing studies the coordination of computers, processors, and/or processes that are physically distributed but work towards a common goal. Many of the fundamental issues involved with distributed computing have been thoroughly researched in the past, for example, synchronization, point-to-point communication, deadlock issues, etc. To date, there is a growing need for the development of applications that can effectively utilize the underlying architecture to solve complex distributed optimization problems. To this end, one can either create a new algorithm specifically for the architecture or modify existing techniques to run on the new architecture. In this work, the latter approach is adopted.
Evolutionary computation (EC) has been shown to be capable of solving complex problems where traditional methods fail to yield satisfactory results. However, to date there has been no research into creating true distributed ECs with distributed genomes. This dissertation presents a set of genetic and evolutionary protocols (GEPs), which are ECs modified to solve distributed problems. To assess their performance of GEPs, we will be testing GEPs on distributed constraint satisfaction problems, where the variables and constraints are geographically distributed among various entities/agents within a distributed system. We will also apply these GEPs to the sensor network tracking problem, and the sensor network sharing problem.||en_US