Phd Candidates

Thesis / Project Title:Model Ensembles for Predictive Drought Severity and Drought Effects Monitoring using Remote Sensing & Socio-Economic data

Thesis / Project Abstract:

The increasing frequency of occurrence of droughts especially in the Greater Horn of Africa and their effects on livelihoods has led to an increase in the demand for ex-ante drought early warning systems that are stable, highly predictive and that have sufficient lead times. The study uses the case study techniques of artificial neural networks (ANN) and support vector regression (SVR) to build predictive models 1-month ahead for both drought severity and effects. Vegetation condition index aggregated over 3 months and nutrition status of children below 5 years are used as the proxy variables for drought severity and effects respectively. Homogenous and heterogenous ensembles are built from the three approaches of simple averaging, ranked weighted averaging and model stacking. We overproduce 244 ANN and SRV models from which we select 111 models for model ensembling. In regression, the stacked heterogeneous model ensemble with an R2 of 0.94 is shown to outperform both the homogeneous ensembles and the individual champion models that post a maximum R2 of 0.83. Similarly, in classification, the heterogeneous stacked ensemble offered a 9 and 11 percentage points’ improvement over the performance of the SVR and ANN champion models respectively with an even better performance in outlier classes. We conclude that despite the computational resource intensiveness of model ensembling, the returns in predictive performance is worth the investment. We nevertheless advise on the building of ensembles from more diverse techniques over extended forecasting periods to estimate the prediction skill of model ensembles over longer lead times.

Thesis / Project  Title: Resource Allocation in TV White Space Network Using a Novel Hybrid Firefly Algorithm 

Thesis / Project  Abstract:

There is continued increased demand for dynamic spectrum access of TV White Spaces (TVWS) due to growing need for wireless broadband. Some of the use cases such as cellular (2G/3G/4G/5G) access to TV white spaces (TVWS) may have a high density of secondary users (SUs) that want to make use of TVWS. When there is a high density of secondary users in a TV white space network, there is possibility of high interference among SUs that exceeds the desired threshold and also harmful interference to primary users. Optimization of resource allocation (power and spectrum allocation) is therefore necessary so as to protect primary users against harmful interference and to reduce the level of interference among secondary users. Existing resource allocation optimization algorithms for a TVWS network ignore interference among SUs, use algorithms that are not computationally efficient with regard to running time or apply greedy algorithms which result in sub-optimal resource allocation.

In the study, an improved resource allocation algorithm based on hybrid firefly algorithm, genetic algorithm and particle swarm optimization (FAGAPSO) has been designed and its performance analyzed for power allocation, spectrum allocation as well as joint power and spectrum allocation. FAGAPSO is a hybrid firefly algorithm that uses final solution of PSO as its initial solution and applies particle swarm optimization concept of pbest and gbest in firefly movement as well as genetic algorithm’s concept of crossover. A continuous optimization version of FAGAPSO has been applied for power allocation while a binary optimization version of FAGAPSO has been applied for spectrum allocation. A binary-continuous optimization version of FAGAPSO has been applied for joint power and spectrum allocation. For joint power and spectrum allocation, firefly algorithm was modified to solve a binary-continuous optimization problem since power allocation is a continuous optimization problem while spectrum allocation is a binary/discrete optimization problem.

Simulation was done using Matlab. The simulation environment in Matlab was developed from scratch.  Cellular network offload to TV white spaces use case was considered. TVWS channels available in Nairobi CBD were considered in the simulation setup. Simulation results show that, compared to firefly algorithm, particle swarm optimization and genetic algorithm, the hybrid algorithm is able to improve the primary user signal to interference noise ratio, secondary users sum throughput and secondary users signal to interference plus noise  ratio in a TV white space network. Only one algorithm considered, Spatial Adaptive Play, has better primary user signal to interference noise ratio, secondary user sum throughput and secondary user signal to interference noise ratio in a TV white space network but it has poor running time.