Theses & Dissertations

Permanent URI for this collection

https://hdl.handle.net/20.500.14170/2002

Browse

Browsing Theses & Dissertations by Issue Date

Jump to a point in the index:
  • Publication
    Numerical simulation of stress shielding induced by crack interaction in human phalanx bone
    ( 2014)
    Siti Aisyah Abdul Halim
    Bone fracture is an injury not uncommon to everyday life. Most of the time, it leaves permanent damage and a long period of recovery. This situation can be prevented if we understand the mechanics and the process of the bone fracture. This study aims is to evaluate stress shielding induced by crack interaction using a simple model based on Linear Elastic Fracture Mechanics (LEFM). This simulation based on the determination of the Stress Intensity Factor (SIF) and the changes of stress shielding in different crack interval towards the human phalanx bone. Numerical simulation had been carried out in this project to understand the stress shielding induced by crack interaction. The results revealed that the interaction of two cracks is directly proportional to the SIF magnitude and interaction factor at the crack tips. The parallel cracks have experienced increasing shielding effect as the cracks interval increase. The crack interaction limit (CIL) and crack unification limit (CUL) also had been accomplished for every range of crack interval in this project. Several improvements will be conducted for future development of this study, including various stresses loading subjected to the model, porous element added in the model, different planes of the model and use various methods in calculating the stress intensity factor (SIF).
      7  4
  • Publication
    Numerical simulation of stress amplification induced by crack interaction in human femur bone
    ( 2014)
    Noor Alia Md Zain
    This research is about numerical simulation using computational method which study on stress amplification induced by crack interaction in human femur bone. Crack in human femur bone usually occur because of large load or stress applied on it. Usually, the fracture takes longer time to heal itself. The crack interaction is still not well understood due to bone complexity. Thus, brittle fracture behaviour of bone may be underestimated and inaccurate. This study aims in investigating the geometrical effect of double co-planar edge cracks on stress intensity factor (K) in femur bone. Other than that, this research also focus in the amplification effect on fracture behaviour of double co-planar edge cracks, where numerical model is developed using computational method. The concept of fracture mechanics and numerical approaches to solve interacting cracks problems using linear elastic fracture mechanics (LEFM) theory is used. So, it is important to study what is the parameter that can minimize the crack propagation to prevent complete failure. This study has shown that the crack interaction limit (CIL) and crack unification limit (CUL) exist in the model developed. In future development of this research, several improvements will be made such as varying the load, applying thickness on the model and also use different theory or method in calculating the stress intensity factor (K).
      15  13
  • Publication
    A study of LIE group method and its application to solve the unsteady transonic flow
    ( 2014)
    Mohammad Huskhazrin Kharuddin
    The non-linear equations of motion describing the unsteady transonic flow in cartesian coordinates are considered in this dissertation. A method known as Lie group which reduce the non-linear partial differential equation to an ordinary differential equation on the basis of the underlying symmetry structure has been used. The Lie method is quite useful in reducing a complex equation to an easy-to-handle ordinary differential equation. By employing the Lie theory, the full one-parameter infinitesimal transformation group leaving the equations of motion invariance is derived along with its associated Lie algebra. Subgroups of the full group are then used to obtain a reduction by one in the number of independent variables in the system. These reductions are continued until an ordinary differential equation is reached. A series type exact solution of these reduced ordinary differential equation is obtained which leads to a series type exact solution of the unsteady transonic flow equation. The Lie group method seems to be an appropriate choice to handle these nonlinear equation.
      4  6
  • Publication
    Monitoring product quality in cliching process: comparison between univariate and multivariate control charts
    ( 2014)
    Ibrahim Sulong
    In this dissertation, we investigate the stability of clinching process for assembly of LED Lamps II. Currently, univariate control charts such as X chart and R chart are develop for monitoring process target and process variability. However, the company management under study is highlighting the issue of quality in a clinching process because univariate control charts allow only monitoring one quality characteritic at a time. In multivariate environment, there are more than one quality characteristics to be monitored and those characteristics are correlated.
      7  9
  • Publication
    Magnetohydrodynamic (MHD) stagnation point flow and heat transfer towards a stretching sheet with suction in a Nanofluid
    ( 2014)
    Nurul Najwa Mansor
    The magnetohydrodynamic (MHD) stagnation point flow and heat transfer towards a stretching sheet in a nanofluid with suction is investigated. The basic equations governing the fluid flow and heat transfer are transformed into nonlinear ordinary differential equation using a similarity transformation, before being solved by means of a shooting method. The effects of suction parameter on the skin friction coefficient, the local Nusselt number and the local Sherwood number as well as the velocity and the temperature profiles are obtained and analyzed. The numerical results indicate that the magnitude skin friction coefficient, the local Nusselt number and the local Sherwood number increase with suction effect.
      2  6
  • Publication
    Modelling and simulation of mixed quieuing server pharmacy system with adjustable parameters
    ( 2015)
    Rashida Abd Rashid
    This dissertation involves discrete event simulation (DES) as a type of computer based modelling that imitates a real world system of pharmacy unit. Queuing theory that is used to model and analyse the characteristic of queuing system at the pharmacy unit of Hospital Tuanku Fauziah, Kangar. The input of this model is based on statistical data collected for 20 working days in June 2014. Currently patient waiting time of pharmacy unit is more than 15 minutes. The actual operation of the pharmacy unit is a mixed queuing server with M / M / 2 queuing model where the adjustable parameter is refer to the pharmacist. Discrete event simulation method and ProModel software is used to simulate the queuing model and to propose the improvement for queuing system at pharmacy unit. Waiting time for each counter is analysed and found out that Counter 3 and 4 has the highest waiting time which is 16.98 and 16.73 minutes. Three scenarios; M / M / 3 , M / M / 4 and M / M / 5 are simulated by using ‘What If Analysis’ and waiting time for actual queuing model and experimental queuing model are compared. The simulation results show that by adding the server (pharmacist) will reduce patient waiting time. Almost 50% average patient waiting time is reduced when one pharmacist is added to the counter. However, it is not necessary to fully utilize all counters because eventhough M / M / 4 and M / M / 5 produced more reduction in patient waiting time, but it is ineffective since Counter 5 is rarely used and current average patient waiting time for Counter 6 is 1.99 minutes
      6  2
  • Publication
    Extended hamming distance method for measuring the perfomance of academic staff
    ( 2015)
    Rasyida Md Saad
    This research presents an extended Hamming distance method in measuring the performance of the academic staff. Generally, performance appraisal is responsible in determining the employee works, in which involving in making personnel decisions and employee development. As this process can be used to identify the performance of the employee, a proper method has to be applied in selecting the best employee performance. Multi criteria decision making (MCDM) method is one of the solution for this problem. In this research one of the existing MCDM methods which is a Hamming distance method is extended. Inspired by the Canós algorithm, the extension of the existing algorithm is done. The use of subjective weight and fuzzy linguistic variables are introduced in the newly extended method. The use of criteria weight is essential in evaluating the performance of the alternative. In addition, Shannon’s entropy concept is also applied to determine the most important criteria. Meanwhile, the use of fuzzy linguistic variables will help in dealing with subjective assessment. The extended Hamming distance method is presented in a well-built framework which also depicted in this research. A study case in one of the institute in Universiti Malaysia Perlis, which is Institute of Engineering Mathematics in determining the best academic staff for three years, which are 2010, 2011 and 2012 is presented to validate the extended method. By utilizing the extended method, the most important criteria, the ranking of the academic staff and problems regarding the ranking process for three years are identified. The comparison between the final results of the extended method and TOPSIS method is also done. Based on the final results obtained, it can be concluded that the results for the criteria weight and ranking of the academic staff for the years 2010 and 2011 are different than the year 2012. For the years 2010 and 2011, A22 is declared as the best academic staff. Meanwhile, for the year 2012, the best academic staff is belong to A3. C11 is declared as the most important criteria for the years 2010 and 2011. However, for the year 2012, C6 is declared as the most important criteria. The similarity in the final results between the extended method and TOPSIS method is also identified. Through the application of the extended method it is hope to ease the difficulty in decision making process especially in determining the best performance of the academic staff.
      18  4
  • Publication
    Heuristic method for the modified model of capacitated arc routing problems
    ( 2015)
    Muhamad Hafiz Masran
    The management of waste collection in urban areas is a critical issue for some developing countries especially in Malaysia which requires immediate attention. This study attempts to make an analysis and assessment of the development of the Perlis state waste generation distribution and the optimization of waste collection routes. This thesis describes the architecture design of the mathematical modeling for the mixed graph of Capacitated Arc Routing Problem (MCARP) extended from the Capacitated Arc Routing Problem (CARP). Furthermore, it provides a description of several equations that can be used to obtain efficient collection routes such as capacity vehicles constraint. To get the better result, this study used the existing heuristics method for solving MCARP mathematical modeling. The result shows that the total costs produced are considerably acceptable and are also reasonably improved, although the existing heuristics method do not obtain an optimal solution but in term of the computational time is good and better than other methods for small and intermediate networks especially in the mixed graph problem
      15  6
  • Publication
    Application of program evaluation and review technique (PERT) and linear programming (LP) in project management
    ( 2015)
    Siti Fairuz A. Rahim
    Throughout the years, businesses execute a systematic change, such as constructing a building, merging with another company or making a new product. These changes lead to the management of a project. To construct a project, various dataflow are required. However, project management nowadays demand more complex products, higher quality outputs and faster development cycles. This will cause massive scale of planning a project which leads to massive cost. Therefore, this study aims for an optimization in managing a project. The Program Evaluation and Review Technique (PERT) which originally designed to plan a project that emphasizes the relationship between the times each activity takes and the resulting time and cost for the expected completion of the entire project. In PERT, a graphical network is used to illustrate the precedence or parallel relationship among the activities. It‟s also identifying the expected completion time of a project and the critical path where more resources should be concentrated to finish the project on time. However, the expected time of a project can also be shorten to a certain level by the used of additional resources which is called crashing. Reducing the duration might include using overtime, assigning more labour and using more resources such as material, equipment and so on. These actions may result in higher cost and lower quality. Thus, crashing decision can be formulated by using Linear Programming (LP). LP technique seeks to determine the optimization project cost. It is also boast efficiency of computational algorithm for problems with thousands of constraint and variables which analyse by using Excel Solver. In this research, PERT and LP are applied on a large-scale of projects which construct a new building consist of Administration Office and Medan Selera at the Main Campus of University Malaysia Perlis (UniMAP), Mukim Padang Siding, Pauh, Perlis
      6  7
  • Publication
    Application of Program Evaluation and Review Technique (PERT) and Linear Programming (LP) in project management
    ( 2015)
    Siti Fairuz A. Rahim
    Throughout the years, businesses execute a systematic change, such as constructing a building, merging with another company or making a new product. These changes lead to the management of a project. To construct a project, various dataflow are required. However, project management nowadays demand more complex products, higher quality outputs and faster development cycles. This will cause massive scale of planning a project which leads to massive cost. Therefore, this study aims for an optimization in managing a project. The Program Evaluation and Review Technique (PERT) which originally designed to plan a project that emphasizes the relationship between the times each activity takes and the resulting time and cost for the expected completion of the entire project. In PERT, a graphical network is used to illustrate the precedence or parallel relationship among the activities. It‟s also identifying the expected completion time of a project and the critical path where more resources should be concentrated to finish the project on time. However, the expected time of a project can also be shorten to a certain level by the used of additional resources which is called crashing. Reducing the duration might include using overtime, assigning more labour and using more resources such as material, equipment and so on. These actions may result in higher cost and lower quality. Thus, crashing decision can be formulated by using Linear Programming (LP). LP technique seeks to determine the optimization project cost. It is also boast efficiency of computational algorithm for problems with thousands of constraint and variables which analyse by using Excel Solver. In this research, PERT and LP are applied on a large-scale of projects which construct a new building consist of Administration Office and Medan Selera at the Main Campus of University Malaysia Perlis (UniMAP), Mukim Padang Siding, Pauh, Perlis
      15  7
  • Publication
    Factors contributing to air pollution levels using factor analysis
    ( 2015)
    Asmarini Mohamed
    The aims of this study is to determine the major factors and causes of air pollution levels within the four selected Malaysian air monitoring stations. The data were obtained from January 2008 to December 2012. The variables concerned are two meteorological factors (temperature and relative humidity) and four air pollutants (CO, PM10, O3 and NO2). The focus areas comprise of four different localities which are; Perai (S1- industrial), Shah Alam (S2-urban), Seberang Jaya (S3-suburban) and Jerantut (S4- control). Two statistical methods were selected to analyze the datasets which are descriptive statistical analysis and factor analysis. The descriptive statistical analysis showed that Perai recorded the highest concentrations of CO, O3 and NO2 while Seberang Jaya recorded the highest concentration of PM10. The results of the factor analysis showed that there were three factors at Perai, Shah Alam and Jerantut while two factors at Seberang Jaya. However, the combination of the components in each factors are different for every station. Nevertheless, it has a significant combination of the components at Perai where it can be categorized as an Organic Pollution Factor, Meteorological Factor and Fuel Factor. These results present a guideline to establish evaluations appropriate to pollution characteristics and classification systems for air quality monitoring stations in Malaysia.
      5  5
  • Publication
    Variational Monte Carlo study of light nuclei
    ( 2015)
    Khairul Anwar Mohamad Khazali
    An outstanding problem in Variational Monte Carlo (VMC) calculations with realistic interactions like Argonne V18 and Urbana IX three-body interactions is that p-shell nuclei turn out to be grossly under bound as compared to the Green’s Function Monte Carlo (GFMC) calculations. A similar situation exists in Diffusion Monte Carlo calculations with somewhat simplified interactions. In this thesis, we improve upon the VMC calculations by bringing about several variations in the established procedure of performing variational calculations. In the first variation, the effect of the errors as a function of the number of particles in the variational wave function are analyzed and then a correction through expanding the radial part in terms of a complete set are made and treat the expansion coefficients as variational parameters. Second variation consists in modifying the variational wave function structure. The state of the art variational wave function for s- and p-shell nuclei consists of two parts, where the first part is a Jastrow part operated upon by a symmetrized sum of two-body operatorial correlations and in the second part this outcome is then operated by a sum of unity, operatorial threebody and spin-orbit two-body correlations. A considerable improvement is obtained over the binding energies, wave functions and variance for the light nuclei 3H, 4He and 6Li by using these two variations. We obtain noticeable improvement in the quality of the wave function and lowering of the energies compared to earlier results. The new energies are –8.38 MeV, –28.07 MeV and –29.90 MeV for 3H, 4He, and 6Li respectively. All the computations have been taken away on a multiprocessor machine developed indigenously
      4  13
  • Publication
    An extended thomas-fermi model with clustering in Nuclei
    ( 2016)
    Nooraihan Abdullah
    In this study, a phenomenological theory of nuclei that includes clustering at the nuclear surface in a general form is presented. The theory elaborates the recently extracted large symmetry energy by Natowitz et al., at low densities of nuclear matter and is fully consistent with the static properties of nuclei. Clusters of all sizes and shapes together with medium alterations are included in a phenomenological way. Discussion about the symmetric nuclear matter properties are presented in detail. Arguments are stated, which bring to an equation of state of nuclear matter consistent with clustering in the low-density region. In addition, the justification and discussion about the properties of asymmetric nuclear matter are given. Due to the clustering, an interesting interpretation of the equation of state of asymmetric nuclear matter comes into view. An extended version of Thomas-Fermi theory is adopted as a framework for nuclei where it also consist of the phenomenological pairing and Wigner contributions. This theory relates the nuclear matter equation of state, which includes clustering at low densities, with clustering in nuclei at the nuclear surface. Calculations are carried out for different equations of state of nuclear matter. This study take into account binding energies of 2149 nuclei for N,Z  8. The significance of the quartic term in symmetry energy is illustrated at and below the saturation density of nuclear matter. It is demonstrated that it is greatly connected to the use of ab initio, a realistic equation of state of neutron matter, especially the contribution originating from the three neutron interactions and rather clustering. The causes for these are presented. The neutron skin thickness in nuclei is found to reduce significantly, for the reason of clustering. The developed theory predicts circumstances and methodologies to be examine both theoretically and experimentally.
      7  6
  • Publication
    A rational cubic spline technique for preserving the positivity of data
    ( 2016)
    Anas Khalaf Faris Al-Abbasi
    This work intends to address the problem of the visualization of curves (2-dimensional data representation) and surfaces (3-dimensional data representation) with the aim and provision that their display looks smooth and modifiable. In order to achieve these goals, we proposed a C1 spline interpolation. For the treatment of the 2D data visualization problem, the proposed function has been made to contain three positive shape parameters in each subinterval of its construction. Simple data-dependent constraints are derived for single shape parameter to ensure preserving the positivity through given positive data while the remaining two parameters are left free for designer’s choice for the curves’ refinement and/or manipulation. This interpolation has been extended into rational bi-cubic spline interpolation to treat the problem of 3D data visualization. The extended interpolation has been made to involve six positive shape parameters in each rectangular patch of the surface construction. In this case, constraints are derived for two shape parameters for conserving the surface positivity while the remaining four parameters are left free according to designer’s will for the surface smoothing and/or manipulating. Half of these shape parameters are set in the x-direction and the other half are set in the y-direction in such a way that each one of the free parameters can be changed separately to obtain different data representation models accordingly. The scheme under discussion is locally effective on the data intervals and does not allow to be inserted with any new knots to preserve the positivity. Numerical examples are provided to demonstrate that the proposed scheme is successfully producing interactive, smooth and modifiable curves and surfaces.
      7  12
  • Publication
    A study in the theory of geometric functions of a complex variable
    ( 2016)
    Hiba Fawzi Sabba Al-Janaby
    This thesis deals with various types of analytic geometric functions in the open unit disk, such as normalized, meromorphic, p-valent, harmonic, and fractional analytic functions. Five problems are discussed. First, the class of analytic functions of fractional power is suggested and used to define a generalized fractional differential operator, which corresponds to the Srivastava–Owa operator. The upper and lower bounds for fractional analytic functions containing this operator are discussed by employing the first-order subordination and superordination. Coefficient bounds for the new subclass of multivalent ( p-valent) analytic functions containing a certain linear operator are then presented. Other geometric properties of this class are studied. A new subclass of meromorphic valent functions defined by subordination and convolution is also established, and some of its geometric properties are studied. For a normalized function, the extended Gauss hypergeometric functions, which are generalized integral operators involving the Noor integral operator, are posed and examined. New subclasses of analytic functions containing the generalized integral operator are defined and established. In addition, some sandwich results are obtained. Third-order differential subordination outcomes for the linear operator convoluting the fractional integral operator with the incomplete beta function related to the Gauss hypergeometric function, are investigated. The dual concept of the third-order differential superordination is also considered to obtain third-order differential sandwich-type outcomes. Results are acquired by determining the appropriate classes of admissible functions for third-order differential functions. The final phase of this dissertation introduces two subclasses of S'h , which are denoted by LH(r) and H(a,B) . Coefficient bounds, extreme points, convolution, convex combinations, and closure under an integral operator are investigated for harmonic univalent functions in the subclasses H(a,B) and Lh (r) . Connections between harmonic univalent and hypergeometric functions are also fully investigated
      8  4
  • Publication
    Approximating one-dimensional coupled shallow water equation for predicting Tsunami wave propagation using new semi-implicit method
    ( 2016)
    Siti Maryam Hafiza Mohd Kanafiah
    Tsunamis are uncommon events compared to natural hazards. Population growth along the shorelines has increased the statistic of the number of victims affected by the tsunami disaster. In order to help people to be aware about the tsunami disasters, a tsunami wave propagation must be predicted. This study proposed a one dimensional coupled shallow water equation for predicting the tsunami wave propagation. The simulation of numerical modelling was conducted to predict the occurrence of tsunami wave propagation. In order to approximate the shallow water equation, the discretization process was conducted using a new semi-implicit method which is one of the methods in finite difference method. A problem regarding the free surface wave damping problem were used in validating the semiimplicit method. After the validation process, the numerical algorithm was developed in MATLAB software to solve the one dimensional coupled shallow water equation for the tsunami wave propagation. For the purposes of the simulation, the boundary condition and initial condition, the spatial steps and time steps as well as the approximation of shallow water equation were encoded. The simulation results of one dimensional shallow water equation successfully predicted the behaviour of the tsunami propagation near the coastline after the earthquake ground shaking occurred.
      5  12
  • Publication
    Prediction of students academic achievement in UniMAP using data mining approach
    ( 2016)
    Fatihah Aziz
    The main objective of this research is to predict the academic achievement of the students in UniMAP who will be succeed, will be continued their study in higher level and also students whose need help in their study. Besides that, the specific objectives of this research are to identify the factors which influence the academic achievement of the students, to develop the decision tree model for prediction of the academic achievement of the students and lastly, to evaluate the performance of the decision tree model. Our research methodology started with the collection of students’ data from UniMAP since 2002 until 2012. Then, continued with the implementation of the decision tree model by using the C4.5 algorithm to develop the model for predicting the students’ academic achievement in UniMAP. C4.5 algorithm is a successor of ID3 algorithm which had been invented by J.R Quilan in 1986. In this research, JMP Pro 11 software has been used in developed the decision tree model. In addition, the finding of this research stated that there were five factor variables had influenced the academic achievement of the students in UniMAP. The factor variables were gender, age, entry qualification, entry CGPA and results from semester 1 and 2. The response variable was the result of graduate students. Moreover, from the result of the decision tree, the knowledge representation has been interpreted into decision rules using IF-THEN rules. From the predictive model, there were 25 decision rules which had interpreted. Later, the rules had been implemented into a simple academic prediction system for the students in UniMAP. This simple prediction system could help new students to predict their final result in the earlier semester. Some of the significant finding from this research, found that students who in between 24 to 30 years old had graduated with excellent result than younger and older adult learners. Female students from this university had graduated with more excellent result than the male students. Lastly, the finding of this research found that students who entry the university with the Matriculation qualification graduated with more excellent result than Diploma and STPM.
      16  13
  • Publication
    Numerical simulation of the burger's equation
    ( 2016)
    Intan Mastura Ramlee
    In this study, the exponential finite difference technique has been used to solve one-dimensional Burgers’ equation with different value of h (step size). Burgers’ equation is considered in this study because the equation governing simple nonlinear diffusion process. Since Burgers’ equation is nonlinear equation, the Hopf-Cole transformation has been applied to convert the equation to one-dimensional heat equation. Consequently, the exponential finite difference method has been used to generate of one-dimensional heat equation. Three techniques called explicit exponential finite difference method, implicit exponential finite difference method and modified Burgers’ equation using explicit exponential finite difference method have been implemented. In the solution process, the explicit exponential finite difference method used a direct to solve the Burgers’ equation while the implicit exponential finite difference method leads to a system of nonlinear equation. At each time-level, Newton’s method is used to solve the nonlinear system. The solution of the onedimensional modified Burgers’ equation by using the explicit exponential finite difference method. The solution process have been discretized the time derivative and spatial derivative using exponential finite difference technique. Numerical solutions for each method are compared with exact solution and the results obtained using three methods are precise and reliable. The percent errors are compute and found to be sufficiently small.
      15  8
  • Publication
    Application of cubic timmer curve with countinuity on batik design
    ( 2017)
    Noor Hairulnisa Noor Mohameed
    In the world of CAGD, to create curve the research tends to use the curve that been usually used such as B-Spline and Bezier. Apparently, the researcher tends to choose cubic curve because it is easier to generate in coding that been created. In this study, Cubic Timmer curve that been chosen to create the curve based on chosen example which is a sample of Batik Design because advantages of Cubic Timmer curve is it tend to go through the midpoint of control polygon that brings a lot of help in creating the curve. Cubic Timmer curves that been created had C0 or C1 continuity by using Mathematica and MATLAB software. After a sample of Batik design is chosen, the sample been inserted into Mathematica then control points are selected. The control points that been selected are insert into coding in MATLAB then been generated to have a curve. Among the curves that been generated, the control point must be adjusted to make the curve having 1 C continuity if the curves are not to be connected by 1 C continuity then it is 0 C continuity. If the curve that been generated is not fulfilled the curve that desired then the control points must be altered to have a good resemble curve as sample. The result is mimicking the sample of Batik design perfectly since of its uniqueness properties that help a lot in creating the curve.
      32  13
  • Publication
    Fuzzy Sumudu transform
    ( 2017)
    Norazrizal Aswad Abdul Rahman
    Ordinary differential Equations (ODEs) have long been used to model real-life problems. As times goes on, researchers realized that ODEs are no longer the best modelling tool, since they are inaccurate, especially when dealing with stuff that are surrounded by uncertainties. To deal with this, researchers utilized the fuzzy set theory to propose fuzzy differential equations (FDEs). Unlike ODEs which could not cope with uncertainties at the initial values, FDEs handle this efficiently. However, the methods that can be used to deal with FDEs are very limited, and many of them are still at development stages. In this research, a new analytical method is proposed for solving a class of FDEs, including FDEs of integer order, fuzzy fractional differential equations (FFDEs), fuzzy partial differential equations (FPDEs) and system of linear fuzzy differential equations (SLFDEs). For this purpose, a novel integral transform called fuzzy Sumudu transform (FST) is proposed. This is done by integrating the fuzzy set theory and the classical Sumudu transform (CST). New properties and fundamental results concerning FST are proposed and proved mathematically. These new properties comprised results on duality with fuzzy Laplace transform (FLT), linearity, preserving, shifting, convolution, as well as theorems for fuzzy derivatives. The proposed results are then used to construct detailed procedures for solving FDEs, FFDEs, FPDEs and SLFDEs where the class of FDEs are interpreted under the strongly generalized differentiability concept. From there, the procedures are demonstrated on some examples, in order to illustrate FST applicability. Some analyses are performed where the behaviour of the solutions obtained are described in particulars. Furthermore, a comparison between the solutions obtained for the class of FDEs are compared with the solutions obtained using ODEs, where the ODEs include integer order ODEs, fractional differential equations (FrDEs), partial differential equations (PDEs) and system of linear differential equations (SLDEs). The main advantage of FST highlighted in this thesis is that it possessed the scale preserving property. This means that the transformed function, with a new domain, is a similitude to the original function. Additionally, this research provides an alternative for researchers when dealing with the class of FDEs.
      6  8