Abstract
Multi-hop Wireless Networks (MHWNs) provide an inexpensive and ubiquitous communication paradigm for the Internet of Things (IoT) to seamlessly network the surrounding objects. However, in order to meet different system requirements such as end-to-end delay and energy efficiency, there are still many challenges that need to be solved in such resource constrained networks. Intelligent task allocation and scheduling offers a promising solution, which enables in-network parallel processing and allows network nodes to share resources and interact with each other via multi-hop communication links. The objective of this thesis is to enhance resource sharing in MHWNs and to improve system performance from the task allocation point of view. This task allocation and scheduling problem in MHWNs is very challenging when complex network aspects are taken into account, such as shared multi-hop wireless communication environment, node heterogeneity and structure-less network topology. Moreover, network dynamics further complicate the problem. For instance, node mobility and failure events can easily cause an optimized task allocation solution to become invalid. In such case, a complete re-run of the optimization algorithm from the scratch is not computationally efficient. To tackle these challenges, this thesis recommends a Genetic Algorithm (GA) based evolutionary task allocation approach and provides an effective, adaptive, and scalable task allocation system for MHWNs. There are three main system components proposed in this thesis: an Intelligent Task Allocation and Scheduling (ITAS) scheme, a Dynamic Task Allocation and Scheduling (DTAS) framework, and a distributed Energy-efficient Clustering (EC) algorithm. ITAS provides suitable task allocation solutions which are able to balance energy consumption while also reducing task processing time and end-to-end communication delay in complex multi-processor and multi-hop wireless networks. Automated task reallocation in a dynamic and mobile environment is performed by DTAS. Finally, EC is beneficial in large-scale networks, offering a scalable and easily manageable hierarchical task allocation structure. By combining ITAS, DTAS, and EC, the proposed task allocation system is able to support high performance applications in resource-constrained networks by assigning suitable tasks to the right resources based on user requirements and cost constraints. It provides optimized task allocation solutions for multi-objective task assignment problems (particularly in complex and dynamic multi-hop wireless environments), which is a clear step forward towards realizing resource sharing in the future IoT systems. Key words: Multi-hop wireless network, Task allocation and scheduling, Genetic algorithm, Network dynamicity, Clustering.