EPSRC DTP PhD studentship: Evolutionary Multi-Objective Workflow Scheduling in Cloud
University of Exeter - College of Engineering, Mathematics and Physical Sciences
|Funding for:||UK Students, EU Students|
|Funding amount:||£14,296 per annum|
|Placed on:||1st November 2016|
|Closes:||11th January 2017|
|★ View Employer Profile|
Main supervisor: Dr. Ke Li (University of Exeter)
Co-supervisor: Prof. Geyong Min (University of Exeter)
Cloud computing is a type of Internet-based computing platform that executes large-scale applications with enormous computational resources to offer on demand. At the dawn of “Industry 4.0”, scheduling for big data and cyber-physical cloud computing has become enormously important. It is well known that the scheduling of a large number of task workflow in a distributed computing platform is a NP-hard problem. The problem becomes even more challenging when executing a large number of tasks on various virtual machines under a cloud environment. One of the major difficulties is the existence of multiple conflicting objectives. For example, a cloud provider may consider minimizing the resource consumption via efficient scheduling, while cloud users concern about the Quality of Service (QoS). Even among the cloud users themselves, different users or even the same user at different time may have different QoS specifications, e.g., low computational costs, high speed.
Due to the population-based property, evolutionary algorithm (EA) has been widely accepted as a major tool for multi-objective optimization. One of the other major merits of EA is that it works without strong mathematical assumption of the problem at hand.
This projects aims at developing effective and efficient evolutionary multi-objective optimization techniques for tackling workflow scheduling problem on an Infrastructure as a Service (IaaS) platform. In particular, it involves taking advantage of the domain knowledge for the development of novel solution encoding schemes, population initialization strategies, fitness assignment methods and genetic offspring generation operators. Moreover, it also involves the development of effective strategies for handling problems with a large number of objectives.
Share this PhD
Type / Role:
South West England