Applied Mathematics and Computational Intelligence (AMCI)

A Deep Learning Approach for Face Detection and Recognition to Initiate Human-Robot Conversation

Tue, 04 Jun 2024 00:00:00 +0000

Artificial Intelligence (AI) is currently booming at almost all field. The inauguration of OpenAI ChatGPT using Natural Language Processing (NLP) has played a vital role in exposing AI to the public. It is estimated about 1.8 billion users visit ChatGPT site in a month, with further planning of apps creation in iTunes Apple App Store and Android Google Playstore. Therefore, it is interesting and natural to implement such technology in robotic field. This paper presents the attempt to employ AI into the mobile robot system towards the main goals of conversational intelligence between human and robot. First, the robot head is designed and assembled, then a screen that functioned as the robot face is attached. Afterwards the detection and recognition system were developed giving the ability to the robot to recognize registered persons and the robot eye is able to track where the person is, in the camera Field-of-View (FOV). In addition, all these systems are developed on in-situ device i.e. NVIDIA® Jetson^TM Nano. It is targeted that the proposed system is able to initiate a natural conversation between a robot and a human user.

Estimation of New Resource Allocation in Hospital’s (or Medical Care) Inpatient Department using Discrete Event Simulation

Tue, 04 Jun 2024 00:00:00 +0000

This study presents a computer simulation model for the inpatient department of a public hospital located in Kelantan by using the Discrete Event Simulation approach. Overcrowding, long wait times, and shortages of nurses and beds have been identified as significant issues in the department, which requires a reliable tool for analyzing current operations and optimizing resource allocation to improve service quality. By applying Discrete Event Simulation, the study models the system of the inpatient department and identifies bottlenecks in the process. Arena software was utilized to determine the optimal number of resources needed to meet the demands. The improvement model was developed based on the optimization results, and its conclusive findings demonstrate that the enhanced model significantly improves the performance of the inpatient department in terms of patient waiting time and the utilization rate of nurses and beds. In addition, a new mathematical equation has been developed to generate alternative resource allocation options based on the hospital budget. The implementation of the new configuration of inpatient department resources, as constructed in this study, effectively improves system bottlenecks. The findings of this study can be used to inform decision-making and enhance the efficiency and effectiveness of the inpatient department.

Development of N-Period Dynamic Programming Model for Determining EOQ under Stochastic Demand

Aisha Sheikh Hassan — Tue, 04 Jun 2024 00:00:00 +0000

This paper demonstrates an approach to determine the EOQ of an item under a periodic review inventory system with stochastic demand. Using N-Period dynamic programming planning horizon, we analyse the ordering policy during each period considering when demand is favorable or unfavorable to determine the EOQ and the associated profit at the end of each planning horizon. The objective is to determine in each period of the planning horizon, an optimal EOQ so that at the long run profits are maximized for a given state of demand. We also generate a formula to determine the number of matrix transitions for each planning horizon until final stage and then prove the formula by the principle of mathematical induction.

A Mathematical Model for The Spread of Varroa-Mites in Honeybee Colony with Fractional Order

Musibau A.Omoloye, Sunday O.Adewale, Saheed Ajao — Tue, 04 Jun 2024 00:00:00 +0000

Honeybees live in colonies with one queen running the whole hive. Worker honeybees are all females and are the only bees most people ever see flying around outside of the hive. They forage for food, build the honeycombs, and protect the hive. This study developed and analyzed the honeybee’s transmission dynamics under fractional order derivative via Laplace Adomian Decomposition Method, spread of varroa-mite by analyzing the disease-free equilibrium and global stability of endemic of our formulated model were investigated. Based on the trajectories, it was concluded that the memory index or fractional order could use to control the honeybees infested by varroa-mite carrying virus transmission dynamics.

Augmented Lagrangian Method for Optimal Control of Interconnected Systems

Wah June Leong, Hong Keat Yap, Hong Seng Sim — Tue, 04 Jun 2024 00:00:00 +0000

This paper explores the application of the augmented Lagrangian method (ALM) for constructing optimal control of some interconnected systems. The ALM proves to be a robust technique in handling the stability and observability restrictions arising from interconnections among subsystems. By segregating Lagrange multipliers from the solution process, the method effectively solves the optimal control problems in a simpler unconstrained setting. The proposed approach is substantiated through numerical simulation, demonstrating its efficacy in obtaining optimal control strategy for the interconnected networks of a power grid model.

Mathematical Modeling of Illicit Drug Use Dynamics Examining the Impact of Recycling Recovered Individuals into the Population.

Tue, 04 Jun 2024 00:00:00 +0000

Illicit drug use continues to pose a significant threat to public health and societal well-being. This study aims to develop a comprehensive mathematical model that captures the dynamics of illicit drug use, considering different categories such as susceptible individuals, exposed individuals, drug addicts, and individuals in recovery. The model accounts for the fact that recovered individuals may not develop lifelong resistance to drugs and that those exposed to drug use cannot introduce new individuals into drug usage. Through the mathematical model, we seek to enhance our understanding of the behavior and patterns of illicit drug use, contributing valuable insights to the field of drug-related research. Additionally, the study explores the properties of the proposed model and utilizes the next generation matrix method to calculate the effective reproduction number. Furthermore, a sensitivity analysis is conducted to evaluate the impact of various parameters on the reproduction number, identifying the most effective control measures for mitigating the spread of drug-induced issues. The findings of this study serve as a warning indication to individuals and governments, emphasizing the need to intensify efforts in combating drug abuse. Moreover, the study highlights the importance of considering backward or forward bifurcation phenomena in determining the local stability of the model system at the endemic equilibrium point. This research contributes to the ongoing efforts to understand and address the complex challenges posed by illicit drug use, offering insights that can inform policies and interventions aimed at curbing the spread of drug abuse and promoting overall community well-being.

A Review on Predictive Model for Heart Disease using Wearable Devices Datasets

Nor Azuana Ramli, Mohd Syafiq Asyraf Suhaimi, Noryanti Muhammad — Tue, 04 Jun 2024 00:00:00 +0000

Heart diseases were the number one killer in Malaysia based on the data from the Department of Statistics Malaysia in the previous year. The number of cases has been increasing from 2156 in 2020 to 2693 in 2021. There were lots of studies that had been done to discover the factors that cause heart disease and ways to prevent it. Among the ways to prevent heart disease include analysis on the patients’ historical data, developing predictive modeling involving statistical and machine learning techniques and monitoring health conditions through wearable devices. This paper reviewed the predictive model that had been applied in heart disease prediction by using wearable devices datasets. Artificial neural networks (ANN) have grown in popularity in data mining and machine learning for its ability to classify input data into several categories by detecting hidden connections in the data, which is beneficial in predicting correct classifications. Other approaches, such as Naive Bayes, neural networks, and Decision Tree algorithms, are used to analyze medical data sets to forecast cardiac disease. Based on the degree of accuracy, Naive Bayes looks to be the most successful model for predicting heart disease patients, followed by Neural Network and Decision Trees.

DenseNet201-Based Waste Material Classification Using Transfer Learning Approach

Michael Tang, Kee Chuong Ting, Nur Hidayatullah Rashidi — Tue, 04 Jun 2024 00:00:00 +0000

This paper explores the application of deep learning models in waste material classification, motivated by the urgent need for efficient waste management practices to address environmental sustainability concerns. Drawing parallels with the success of deep learning in healthcare domains, the study investigates the effectiveness of various deep learning architectures for waste material classification. The DenseNet201 model is proposed and compared with various deep learning models such as ResNet, MobileNetV2, AlexNet, and GoogleNet. Experimental results demonstrate that DenseNet201 achieves superior accuracy, average recall, and average precision, making it the most effective model for waste material classification. The dense connectivity and feature aggregation capabilities of DenseNet201 contribute to its outstanding performance, showcasing its potential for enhancing waste management processes.

Modelling the Early Outbreak of Covid-19 Disease in Malaysia Using SIRS Model with 4-Step Adams-Bashforth-Moulton Predictor-Corrector Method

Abdulwaheed Salaudeen, Associate Professor Saratha, Mr. Cheah, Tan Yi Hang — Tue, 04 Jun 2024 00:00:00 +0000

Compartmental models have gained tremendous usage in various fields, including epidemiology, pharmacokinetics, and ecology, to describe the dynamics of a system consisting of interacting compartments one need to understand that the challenges in using compartmental models is solving the system of ordinary differential equations (ODEs) that govern the dynamics of the compartments. In this study, we propose the use of the Adam-Bashforth predictor method to solve compartmental models formed from COVID-19 data in Malaysia between a specified period and we showcased its promising results. The Adam-Bashforth predictor method is a widely used numerical method for solving ODEs. It uses previous solution values to calculate the next solution value, and the solutions is refined by using another fashion of the formula known as ABM correction formula. We improved the performance of the Adam-Bashforth predictor method by using the first four solutions of the fourth-order Runge-Kutta method, which is another popular numerical method for solving ODEs, using the compartmental models. Our results showed that the Adam-Bashforth predictor method enhanced with the fourth-order Runge-Kutta method for accuracy and computational efficiency was able to capture the trend in the COVID-19 dataset used. Generally, the Adam-Bashforth method was about 2.5 times faster than the fourth-order Runge-Kutta method while maintaining similar accuracy. So merging two of them will in no doubt provide a better accuracy in solving the epidemiological model, the Adam-Bashforth method showed significantly accuracy, particularly in the early stages of the outbreak. The Adam-Bashforth predictor method is a promising numerical method for solving compartmental models. It offers better accuracy and computational efficiency, can be particularly useful in scenarios where accurate and fast predictions of compartmental dynamics are crucial. The model will be of great importance to the Malaysian Government, the Ministry of Health, and other stakeholders in disease management for an immediate and timely response to disease outbreaks.

Eigenvalue Elasticity Analysis of Mathematical Model of Dynamics of Diabetes and its Complications in a Population

Aye Patrick Olabanji — Tue, 04 Jun 2024 00:00:00 +0000

In this paper, a mathematical model of dynamics of diabetes and its complications was presented to explore the parameters with the greatest impact on the model. The model allows for the individuals to move from the susceptible class to the treated class. The model exhibit one equilibrum state, namely, the disease prevalent equilibrium state. The local and global asymptotic stability of the equilibrium state was determined using quadratic Lyapunov method. Eigenvalue elasticity analysis of the model parameters was carried out and parameter d (mortality rate due to complications) has the highest positive eigenvalue elasticity value. Also, using the eigenvalue sensitivity analysis, the parameter d has the highest positive value. The analysis revealed that parameter d has the greatest impact on the formulated mathematical model of disease dynamics which must be put into consideration by the health care policy makers in order to reduce the rate of mortality due to the disease.