http://hdl.handle.net/123456789/11 2026-04-09T00:56:34Z 2026-04-09T00:56:34Z JOSHUA, M. B. http://hdl.handle.net/123456789/4527 2025-11-24T11:39:53Z 2025-01-01T00:00:00Z

Title: MULTI-OBJECTIVE OPTIMIZATION MODEL FOR FARM PLANNING IN SOME SELECTED SMALL-SCALE FARMS IN NAVRONGO, GHANA Authors: JOSHUA, M. B. Abstract: The small-scale agricultural sector makes a substantial contribution to a country's economic growth. The goal of this research is to create a multi-objective optimization model for farm planning that maximize returns whilst minimizing labour cost for optimal land use. To do this, three optimization models were developed using the weighted sum and epsilon-constraint method for multi-objective programming and solved by optimization techniques, using a management scientist software version 6.0. These three models were discussed in detailed and tested using data collected from some selected small-scale vegetables, cereals and legume farmers in some parts of Navrongo in the Kesena Nankana Municipality of the Upper East region of Ghana. Model 1 suggested an optimal land use and an increased in returns for all the test it was employed on. Also model 2 showed an increased in returns in all test cases while suggesting an optimal land use as well. Furthermore, model 3 showed a decreased in cost of employing labour whiles suggesting an optimal land usage. In all cases, the models were found to be robust for both the vegetables, cereal and legume farming through the sensitivity analysis done for the range of values for the coefficients of the decision variables and the constraints. In general, the developed models helped the small-scale farmers to maximize returns as well as minimizing the cost of labour whilst proposing optimal use of the farm land. Consequently, it is recommended that future works could consider adding other parameters like fertilizers application or type of fertilizer used, soil type, water requirements, etc. to see how the model will work. Description: AWARD OF MASTER OF PHILOSOPHY IN MATHEMATICS

2025-01-01T00:00:00Z Biney, G. http://hdl.handle.net/123456789/4019 2023-05-26T10:38:19Z 2021-01-01T00:00:00Z

Title: ROBUST ADAPTIVE SCHEME FOR GAUSS MARKOV MODEL Authors: Biney, G. Abstract: The Hogg’s adaptive scheme is extended to the Gauss Markov Model. The Gauss Markov model is a statistical procedure which belongs to the class of general linear model. Gauss Markov model is very sensitive to nonnormality, variance heterogeneity as well as large sample size. These assumptions may be violated as a result of departures from normality and small sample size. To overcome these problems, an Adaptive Scheme is adopted. The Adaptive Scheme is a two step procedure in which a selector statistic is used to first exam ine and classify given data based on measures of skewness and tailweight. Afterwards, a test statistic, independent of the selector statistic is chosen and a test conducted. A One way Analysis of Variance and Repeated Measures Design models were considered under uncorrelated and correlated error distributions respectively. The nine winsorised scores proposed by Hettmansperger (1984) were used because they are considered the most ap propriate rank scores for hypothesis testing. The Winsorised scores as well accommodate a wide range of distributions which are either symmetric or asymmetric with varying tail weights. In addition, the benchmarks for cut-off values for the measures of skewness and tailweights postulated by Al-Shomrani (2003) in his PhD dissertation were used. 10,000 simulations were conducted to compare the performance of the Adaptive Scheme and the Gauss Markov model from different continuous distributions under uncorrelated and correlated errors. Analyses of real datasets were as well performed to ascertain the effi ciency of the two tests. The findings favoured the Adaptive Scheme under a broad class of continuous distributions especially for non-normal distributions. The adaptive scheme is applicable to both small and large samples. It is therefore recommended that Statis ticians, Researchers and Data Analysts be encouraged to use adaptive schemes because they are applicable to a broad class of distributions. Description: DOCTOR OF PHILOSOPHY IN APPLIED STATISTICS

2021-01-01T00:00:00Z Yellakuor, B. E. http://hdl.handle.net/123456789/3678 2022-07-28T11:40:54Z 2016-01-01T00:00:00Z

Title: THE THEORY AND KEY TECHNOLOGY RESEARCH IN SOCIAL NETWORKS Authors: Yellakuor, B. E. Abstract: Currently social networks are the pulse of humanity. For there are the platforms on which people share content and form links among themselves for various purposes The traditional social networks that were formed by human relationships or activities have been supplanted with the online social networks (OSNs). And due to the availability of affordable and portable digital devices and also the emergence of the Internet and services offered by the Web 2.0, there have been a tremendous amount of data generated from the digital human activities on these sites. These digital human activities become complex in the process of time and thus practically infeasible to understand the nature of their underlying wiring (topology) and properties via observation or manual means. Due to this various algorithms have been developed and research done to help understand the network topology and properties of social networks and there have been some good findings. But there are still major challenges associated with the theory and key technology research in the field of social network study which this dissertation addresses as follows Firstly the lack of a comprehensive and a fi rm theoretical framework for the in-depth study and analysis of social network is a concern in the research community. We address this challenge by outlining a thorough and well-balanced presentation of the key network theoretical framework and used it to carry out some novel network-centric analysis on the social and other complex networks studied in this dissertation. Secondly, we develop algorithms and use the network theoretical tools outlined in this thesis to analyze the structural and spectral properties of social networks Visual presentations of the correlation of their structural and spectral properties are made and various novel results are presented . These results present a better perspective for modelling real life networks than the one-dimensional methods used in the literature . Thirdly, the properties of social networks are not well applied in other areas of research . But social network structural and spectral properties contain rich data that can be leveraged for modeling real life system. s To this end, we apply the properties of social network studied in this dissertation to model three real life situations. They include; Human disease modeling: Most of the epidemic models in literature used the random network model to simulate the epidemic spreading in human society. However, human society is not random in many aspect , s as such these results do not give a true picture of human disease spreading. Social networks have a high degree of resemblance to human society , as such, we leveraged their properties to model epidemic spreading and reported novel findings •Modeling information spreading or maximization: We develop an information maximization algorithm that scale with the size of the network and use it to model a novel information or influence maximization system and reported results that are better than the state-of-the-art algorithms that used greedy approach. • To infer and measure trust: The graph-theoretical properties of social networks can infer and measure trust more accurately than the current state of the art systems. For instance, OSNs are known to possess a tight core that has a high level of reciprocity among the user. s Users within the core are reached via many short paths, and thus a malicious user would be hardly trusted unless she is able to penetrate the core by skewing many short paths and thus appearing trustworthy to a larger percentage of the core user. s We use these analytical insights to propose a trust metric and use it to measure trust levels on 4 OSN s and reported various novel findings that can serve as a benchmark for inferring trust on an unknown user. The significance of these findings is enormous . These results would be of interest to health and governmental institutions on how they can leverage on the latent amount of data from social networking sites in order to model and study human disease spreading and behaviour patterns. Auction houses and marketing companies can leverage on these, findings to boost their online marketing and auctions strategies . Furthermore, the police and other security organizations can leverage on the finding from this research to enhance their know-how on combating terrorism and financial frauds, through the use of these network theory tools. Description: DOCTOR OF PHILOSOPHY IN COMPUTER SCIENCE AND TECHNOLOGY

2016-01-01T00:00:00Z Abukari, A. M. http://hdl.handle.net/123456789/3535 2022-04-25T11:31:19Z 2022-01-01T00:00:00Z

Title: SECURITY AND STORAGE ENHANCEMENT OF CLOUD ENTERPRISE RESOURCE PLANNING DATA USING HOMOMORPHIC ENCRYPTION AND SECRET SHARING Authors: Abukari, A. M. Abstract: In this thesis, a number of solutions are proposed to enhancing and improving the security and confidentiality of Cloud Enterprise Resource Planning (ERP) Data. Firstly, the Asmuth-Bloom, Blakley, Mignote and other Secret Sharing Schemes (SSS) are reviewed, adopted and modified in order to present a relatively improved secret sharing scheme. Conditions for the scheme is also presented as well as algorithms for implementation of the scheme presented by this research. Secondly, a hybrid of two homomorphic encryption scheme is presented to address chosen ciphertext attacks (CCA) on Cloud ERP Data. The Rivest-Shamir-Adleman (RSA) and Paillier cryptosystems are adopted and modified to present an improved double-layer encryption homomorphically. A System architecture for Video Conferencing in the midst of the pandemic COVID-19 and beyond is presented as well as algorithms for the implementation of same. The hybrid of two homomorphic encryption schemes presented in this thesis do not share keys with the cloud. Thirdly, this thesis presents Homomorphic encryption scheme using the Redundant Residue Number System (RRNS), Geometric Probability, Bernoulli Probability and the concept of secret sharing schemes (SSS). Parameters are deducted and presented based on the reports from Kaspersky lab. The effectiveness of the scheme presented in this research work is demonstrated in the ability to handle data redundancy as well as error detection and correction. Finally, a comprehensive load balancing scheme is presented to handle load management of the Cloud ERP Data shares in a multi-cloud environment. The Weighted Round Robin (WRR) scheme is modified. A dynamic weight (Wd) is introduced to share the Cloud ERP Data shares in the multi-cloud environment. The dynamic weight (Wd) is calculated using the data variance, the root mean square to generate the dynamic coefficient. The load balancing scheme presented solves data loss and delay concerns in load balancing. All the proposed schemes are meant to enhance the security, efficiency and computational integrity of Cloud ERP data homomorphically. The performance of the proposed schemes are evaluated theoretically and simulated using python and compared with other schemes. The comparison analysis suggests the proposed schemes presented in this thesis work offer a substantial improvement over the other schemes. Description: DOCTOR OF PHILOSOPHY IN COMPUTATIONAL MATHEMATICS

2022-01-01T00:00:00Z