Sensor networks feature low-cost sensor devices with wireless network capability, limited transmit power, resource constraints and limited battery energy. The usage of cheap and tiny wireless sensors will allow very large networks to be deployed at a feasible cost to provide a bridge between information systems and the physical world. Such large-scale deployments will require routing protocols that scale to large network sizes in an energy-efficient way. This thesis addresses the design of such network routing methods. A classification of existing routing protocols and the key factors in their design (i.e., hardware, topology, applications) provides the motivation for the new three-tier architecture for heterogeneous networks built upon a generic software framework (GSF). A range of new routing algorithms have hence been developed with the design goals of scalability and energy-efficient performance of network protocols. They are respectively TinyReg - a routing algorithm based on regular-graph theory, TSEP - topological stable election protocol, and GAAC - an evolutionary algorithm based on genetic algorithms and ant colony algorithms. The design principle of our routing algorithms is that shortening the distance between the cluster-heads and the sink in the network, will minimise energy consumption in order to extend the network lifetime, will achieve energy efficiency. Their performance has been evaluated by simulation in an extensive range of scenarios, and compared to existing algorithms. It is shown that the newly proposed algorithms allow long-term continuous data collection in large networks, offering greater network longevity than existing solutions. These results confirm the validity of the GSF as an architectural approach to the deployment of large wireless sensor networks.