His main interest lies in the autonomous decision making using algorithms to help human make decisions such as in negotiations. Particular interest is in the development of Machine Learning algorithm that analyzes data to make informed decisions.
Ongoing / Past Projects
A Self-adaptive Multi-Objective Optimization Algorithm for Dynamic Environment
Solving complex multi-objective optimization (MOO) problems can be painstakingly difficult endeavour considering multiple and conflicting goals. One of the challenges in MOO algorithms is to adjust themselves in a dynamic environment given different goals dimensions (i.e. single/multi-dimensional independent/multi-dimensional dependent). In a changing environment, the current optimization goals may need to be altered at a different time to get an optimized result. The complexity of adapting and tuning the goals is higher for multidimensional goals. This research is aspired to use adaptive decision making to address the limitation of the MOO approaches and improve its performance in the changing environment.
DOUBLE AUCTION-BASED NEGOTIATION FRAMEWORK FOR HETEROGENEOUS AND MULTI-ATTRIBUTES CLOUD SERVICES
Aims at enhancing the double auction framework to accommodate both heterogeneous and multi-attributes cloud services negotiation at the same time. Heterogeneous cloud services are common nowadays where many different sizes and pricing of cloud service virtual machines, processors and storages are supported by cloud service providers. The heterogeneous cloud services require the negotiation frameworks to accommodate multi-attributes negotiation to optimize the benefits to both customer and provider.title
Autonomous Agent-based Negotiation and Dynamic Pricing Framework for Cost and Time Efficiency in Cloud Services Marketplace
The growth of public cloud offerings provide great challenge to consumers in choosing the services, prices and providers that can fulfill their needs. RightScale survey in 2016 identified that the lack of cloud resource/expertise which results from poor management is now the number one cloud issue replacing security. Furthermore, the cloud consumers' top priority is to save costs. To address these issues, the research aims to work at the fundamental level by proposing a multilateral agent based negotiation framework and protocol for cloud service market that is able to regulate dynamic pricing based on the current market condition. The framework could be later implemented for managing large bargaining scenarios in online cloud marketplace and expected to be time and cost effective.
Concurrent Service Level Agreement Negotiation in Cloud-based Systems
The research aimed to address missing components of concurrent service level agreement negotiation in cloud-based systems, namely dynamic pricing and concurrent negotiations. We had identify suitable protocols for concurrent cloud service negotiation and proposed an enhanced concurrent cloud service negotiation mechanism that adopts dynamic pricing for cloud service negotiation. The mechanism not only considers the business objectives, but also account for the precise service level values for better service delivery. This enables customers to acquire the best quality of service that meets their needs. On the other hand, the proposed concurrent negotiation protocol could perform better in terms of negotiation time, number of proposals and average utility in a large market involving multiple buyers and sellers. Hence, the concurrent negotiation protocols are relevant for the cloud service market.
CHILDHOOD OBESITY AND OVERWEIGHT PREDICTIONS USING HYBRID DATA MINING TECHNIQUES
Childhood obesity has become a worrying global epidemic. Evidences show that childhood obesity persists into adulthood. Therefore, predicting obesity at an early age is both useful and important. The objectives of this study are to identify significant risk factors and protective factors of childhood obesity in Malaysia; to select and propose parameters for prediction; to investigate suitable data mining techniques for childhood obesity and overweight predictions; and to propose hybrid data mining techniques to increase the sensitivity of predictions.