E-Services Barriers in Developing Countries

Paper ID : O007 Solving the Goods Transportation Problem Using Genetic Algorithm

Amirah Rahman, Nazmi Syazwan Shahruddin & Ismail Ishak

School of Mathematical Sciences, Universiti Sains Malaysia, 11800 USM, Penang, Malaysia Goods transportation is a critical part of supply chain management and the distance covered in the delivery pro-cess indirectly reflects the sustainability level of the supply chain, especially in the environmental aspect. In the wake of climate change issues faced worldwide and the increasing public concern in pollution reduction, it would be in the best interest of companies to not just minimize their operational costs but to also do their part in reduc-ing externalities such as air pollution and noise pollution. Applyreduc-ing environmentally-friendly solutions to their goods transportation system would bring benefit in terms of both operational cost reduction and increase in brand reputation. In this study, a goods transportation problem is tackled to reduce the distance covered in the trans-portation process by modelling the problem as a Travelling Salesman Problem (TSP) and solving it using Genetic Algorithm. We propose a new crossover operator namely the Nearest-Node Pairing Crossover (NNPX) that is specifically designed to tackle a Travelling Salesman Problem (TSP) by exploiting the distance aspect of the prob-lem. We evaluate the performance of NNPX compared to two other crossover operators: Order Crossover (OX) and Position-Based Crossover (PBX). The results reveal that the performance of NNPX is outstanding compared to OX and PBX. We found that NNPX has a better rate of convergence as it consistently yields lower distances in fewer iterations. In addition, NNPX does not depend on a large population size for faster convergence. In a nutshell, this study proposes a new crossover operator NNPX that is comparatively more efficient when used to solve the goods transportation problem, thus reducing the associated operational cost and externalities.

Keywords:Goods Transportation Problem; Genetic Algorithm; Travelling Salesman Problem; Order Crossover;

Position-based Crossover; Nearest-node Pairing Crossover

Paper ID : O008 Routing Mail Delivery from a Single Depot with Multiple Delivery Agents

Amirah Rahman, Tan Hoong Inn, William Liew Wei Seong & Nazmi Syazwan Shahruddin

1School of Mathematical Sciences, Universiti Sains Malaysia, 11800 USM, Penang, Malaysia Transportation is a crucial part of the mail delivery process and usually associated with environmental sus-tainability problems such as air pollution from emissions of Greenhouse Gasses (GHG). Optimizing the distance covered by delivery agents in the mail delivery process by utilizing the Travelling Salesman Problem (TSP) mod-els and techniques can simultaneously minimize the operations cost and GHG emissions. TSP is a well-known optimization problem that searches for the lowest cost route in network of cities where a salesman must travel through all cities exactly once and returns to the starting city. A Multiple Travelling Salesman Problem (mTSP) is a generalization of the TSP, where there are multiple salesmen instead of one and there exist a base or depot node where all salesmen depart from. All other nodes must be visited only once by one salesman. The aim of this study is to solve a mail delivery problem with multiple delivery agents in an established mail delivery network and compare our optimized delivery route with the original delivery route. This study also aims to optimize the delivery route for different numbers of delivery agents to prepare for cases of employee shortage and vehicle breakdown. This study decomposes the mTSP into multiple separate TSPs by clustering nodes together using the k−Means Clustering method and assigning a delivery agent to each cluster, subsequently solving each cluster of TSP using Genetic Algorithm. The results show that our method is able to provide a better route for the mail delivery system compared to the original route, with a reduction of 11.25% in distance, which would no doubt reduce the transportationrelated GHG emissions.

Keywords:Mail Delivery Problem; Travelling Salesman Problem; k-means Clustering; Genetic Algorithm

Paper ID : O009 Estimation of Maximum Sustainable Yield (MSY) for Sustainable Fish Catch

Nor Atiqah Jamaluddin, Siti Aida Sheikh Hussin, Zalina Zahid & Siti Shaliza Mohd Khairi Faculty of Computer and Mathematical Sciences, Universiti Teknologi MARA, 40450 Shah Alam, Selangor,

Malaysia

Fishery is an entity engage with fish catch on the sea ground by fisherman. The fish catch in Peninsular Malaysia has declined by more than 90% since 1960s. Thus, it is important to know the maximum sustainable yield in order to overcome the decreasing number of fish and estimate the number of fishes that can be caught from the sea without endangering the fish population. The main purpose of this study is to compare the maximum sustainable yield of fish catch in Terengganu by comparing linear regression and Bayesian surplus method. The results show linear regression performs better in estimating maximum sustainable yield at Terengganu as compare to Bayesian surplus model with higher adjusted R-square value and smaller AIC value. The amount of maximum sustainable yield using linear regression is 102983.05 tones and implied 237288.1356 trips as the optimum number of fishing trip per year. Both values are higher than Bayesian surplus model that only allows 100100 tons of maximum sustainable yields and 205700 trips per year. Based on the value of maximum sustainable yield and optimum number of fishing trip from both methods, the current fish stocks have been overfished from year 2000 to 2008 since the fish catch and fishing effort for the year exceeded the optimum value. As a result, the numbers of fish catch for the year 2009 to 2016 was lower even though the fisherman put higher fishing effort.

Keywords:Fish Catch, Maximum Sustainable Yield, Regression, Bayesian Surplus Model

Paper ID : O012 A Bootstrap Simulation for comparison of Group Risk Plan and Multi-Peril Crop Insurance Policy

Valantino Agus Sutomo¹, Dian Kusumaningrum¹, Rahma Anisa²& Aryana Paramita¹

1Prasetiya Mulya University, Jl BSD Raya Utama, BSD City-Tangerang, Prasetiya Mulya University

2nstitut Pertanian Bogor, IPB Campus, Jalan Meranti Wing 22 Level 4, Dramaga, Babakan, Dramaga, Bogor, Institut Pertanian Bogor

Agricultural insurance is one of the solutions for farmers to avoid risks such as climate change, pest, disease and price fall which increase the risk of crop failure. Indonesian Ministry of Agriculture (MoA) and Jasindo have been using subsidized Multi-Perils Crop Insurance (MPCI) and faced some disadvantages that mainly due to high risk moral hazard, adverse selection, and high administrative costs. Our research provides an alternative policy area yield index insurance known as Group Risk Plan (GRP) to handle this problem. To find homogeneous basis risk in GRP, the data was classified into two categories: up to and above two hectares. Bootstrapping approach was used to find MPCI and GRP indemnity amount for each term of harvesting period. In addition, both parametric estimation and goodness of fit test indicated appropriate distribution used by MPCI and GRP in each term of harvesting period as well. To compare the aggregate loss distribution, the conditional value at riskCVaR(X) measures the mean of excess loss exceedingVaR(X). It reflects how risky and measures the amount of money insurance company should retain. Simulations indicated that MPCI has greaterCVaR(X)than GRP, which implies the higher likelihood of loss. The insurance company should have enough reserve for MPCI to cover that loss. On the other hand, the GRP policy that minimizes moral hazard and adverse selection thereby reduces administrative costs and offers potential to market at lower costs, may encourage farmers to do good farming practices that will result in the lower likelihood of loss. Therefore, GRP could be considered as an alternative crop insurance policy in Indonesia

Keywords:Multi-Peril Crop Insurance; Group Research Plan; Bootstrapping; CVaR(X)

Paper ID : O013 Hyperparameters Tuning of Random Forest with Harmony Search in Credit Scoring

Rui Ying Goh¹, Lai Soon Lee^1,2& Mohd Bakri Adam^1,2

1Laboratory of Computational Statistics and Operations Research, Institute for Mathematical Research, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

2Department of Mathematics, Faculty of Science, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

Correct identification of defaulters and non-defaulters in the lending industry is a crucial task for financial insti-tutions. Credit scoring is a tool utilized for credit granting decisions. Recently, Random Forest (RF) is actively researched in credit scoring due to two main benefits i.e. non-parametric flexibility to account for various data patterns with good classification ability and the computed features importance that can explain the attributes. Hy-perparameters tuning is a necessary procedure to ensure good performance of a RF. This paper proposes the use of a metaheuristic, Harmony Search (HS), to form a hybrid HS-RF to conduct hyperparameters tuning. A Modified HS (MHS) is also proposed, forming MHS-RF, for effective yet efficient search of the RF hyperparameters. Along with parallel computing, MHSRF effectively reduces the computational efforts of the hyperparameters tuning pro-cedure. The proposed hybrid models are benchmarked with standard statistical models on the Lending Club peer-to-peer lending dataset. The computational results show that a well-tuned RF have better performance than statistical models, with MHS-RF reported the best performance yet being the most efficient in hyperparameters tuning of RF.

Keywords:Credit Scoring; Random Forest; Harmony Search

Paper ID : O014 Harmony Search Algorithm for Location-Routing Problem in Supply Chain Network Design

Farahanim Misni^2,3& Lai Soon Lee^1,2

1Laboratory of Computational Statistics and Operations Research, Institute for Mathematical Research, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

2Department of Mathematics, Faculty of Science, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia

1Faculty of Industrial Sciences & Technology, Universiti Malaysia Pahang, 26300 Gambang Pahang, Malaysia The interdependence of facility location and vehicle routing has been recognized among the practitioners and researchers. The integration of these problems is more challenging in a supply chain network design prob-lem. This paper considers a location-routing problem, which integrates the facility location problem and the vehicle routing problem. To effectively solved the problem, it is decomposed into three sub problems which are: location-allocation problem, multi-depot vehicle routing problem and multi-depot routing-allocation prob-lem. The objective is to minimize the total operating facilities cost and the total travel distance cost between the depots and customers. A harmony search algorithm is proposed with several local optimisation approaches to further enhance the solution quality. The problem is tested with several benchmark dataset from the literature.

The performance of the proposed algorithm is compared with other heuristic and metaheuristic approaches from the literature.

Keywords:Harmony Search; Location-allocation; Vehicle Routing; Supply Chain

Paper ID : O015 Urban Transit Frequency Setting using Multiple Tabu Search with Parameter Control

Vikneswary Uvaraja¹& Lai Soon Lee^1,2

1Department of Mathematics, Faculty of Science, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia

2Laboratory of Computational Statistics and Operations Research, Institute for Mathematical Research, Universiti Putra Malaysia, 43400 UPM Serdang, Selangor, Malaysia

Urban transit frequency setting is one of the multiobjective problems in public transportation system, which aims to find optimal time interval between subsequent buses along the routes. In this study, a Multiple Tabu Search (MTS) algorithm is employed to determine the bus frequency of the routes that minimize the number of buses, total waiting times and overcrowding simultaneously. The best fit of the algorithm have great influence in the optimization process and the quality of the results. The efficiency of the algorithm is tested on benchmark dataset by changing the value of the total domains. The chosen parameter gives considerable effect on the objective functions compared to other parameters such as the size of tabu list and the number of iterations. Using statistical hypotheses evaluation, significant difference in the performance of the proposed MTS algorithm, in terms of objective function values are measured. The statistical results indicate that the number of domains determines the quality of solutions for different instances of the problem. Additionally, the frequency setting problem is extended by revising the passenger assignment procedure and frequency optimization process with time-dependent demand in order to reflect a real-world situation. The extended results of the algorithm from different size of routes are presented to show the effectiveness of the proposed algorithm.

Keywords:Urban Transit Frequency Optimization; Multiple Tabu Search; Parameter Evaluation; Statistical Analysis

Paper ID : O016 Fuzzy Analytic Hierarchy Process using Intuitive Vectorial Centroid for Eco-friendly Car Selection Fadhilah Che Jamil^1,2, Adam Shariff Adli Aminuddin¹, Ku Muhammad Na’Im Ku Khalif¹& Nor Izzati Jaini¹

1Faculty of Industrial Sciences and Technology, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Kuantan, Pahang, Malaysia

2School of Foundation and Interdisciplinary Studies, DRB-HICOM University of Automotive Malaysia, Peramu Jaya Industrial Area, 26607 Pekan, Pahang, Malaysia

Eco-friendly car is expected to be the next driving market force for global transportation and technology due to its paramount importance towards the sustainability of the environment and society. However, the actual sales of eco-friendly car are not that convincing and it is even decreasing because the consumer is still uncertain to consider eco-friendly as one of the criteria for them to buy their cars. This situation is worsen by the lack of information and awareness regarding sustainability transportation initiatives. Due to the uncertainty and vague understanding of the consumer about this problem, this paper attempts to investigate the current preference of consumer to buy their cars, and whether they really need to buy the eco-friendly car by using the Fuzzy Analytic Hierarchy Process (FAHP) which implements the Intuitive Vectorial Centroid (IVC). Based on FAHP, the imprecise or fuzzy judgment from the decision maker can be incorporated, to anticipate a better decision for eco-friendly car selection. The outcome of FAHP is compared with crisp Analytic Hierarchy Process (AHP), and the findings shows that FAHP can provide an accurate and consistent result with AHP, although it deals with fuzzy judgment inputs from multiple decision makers.

Keywords:Fuzzy Analytic Hierarchy Process; Intuitive Vectorial Centroid; Eco-friendly; Car Selection

Paper ID : O017 A Fundamental Study of Optimizing Back Pain Medicinal Cupping Points Disease via Graph Colouring

Nurfatihah Mohamad Hanafi¹, Yuhani Yusof¹, Mohd Sham Mohamad¹, Muhamad Faiz Abu Bakar²& Mohd Adhha Ibrahim³

1Faculty of Industrial Sciences & Technology, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Kuantan, Pahang, Malaysia

2Faculty of Computer Systems & Software Engineering, Universiti Malaysia Pahang, 26300 Gambang Kuantan, Pahang

3Pusat Bekam Al-Yakin (L0120161637), Kuantan Pahang

Medical cupping is an ancient method that is still used by many people. It is an alternative method in healing process for much type of diseases by using special cups on the specific points of the skin to create suction. Every disease has their specific cupping points. As a habit, practitioners will cup based on patient request or complaint, where it give effect to the cost and time. Thus, this paper is to study the effective ways on finding the optimum number of cupping points for back pain disease through graph colouring. In addition, a formulation of algorithm is given too. The finding shows that the optimum number of back pain medicinal cupping points is two.

Keywords:Medicinal Cupping; Back Pain; Graph Coloring; Mathematical Modeling

Paper ID : O018 A Comparison of Simulated Annealing Cooling Strategies for Redesigning a Warehouse Network Problem

Rozieana Khairuddin¹& Zaitul Marlizawati Zainuddin^2,3

1Faculty of Industrial Sciences & Technology, Universiti Malaysia Pahang, Lebuhraya Tun Razak, 26300 Kuantan, Pahang, Malaysia

2Department of Mathematical Sciences, Faculty of Sciences, Universiti Teknologi Malaysia, 81310 Johor, Malaysia

3UTM Centre for Industrial and Applied Mathematics & Department of Mathematical Sciences, Faculty of Sciences, Universiti Teknologi Malaysia, 81310 Johor, Malaysia

Simulated annealing (SA) has been a very useful stochastic method for solving problems of multidimensional global optimization that ensures convergence to a global optimum. This paper describes the use of SA for solving the warehouse redesigning network problem and compares the performance of three different SA cooling sched-ules: the basic geometric cooling schedule, logarithmic and linear. Extensive computational results, which are performed and described, show that the geometric cooling schedule produces consistently better quality solutions in less time than the solutions produced by the other schemes.

Keywords:Simulated Annealing, Redesigning Network Problem, Cooling Schedules

Paper ID : O020 The Effects of Fuel Price Fluctuation on Household Income in Malaysia

Norhana Abd. Rahim & Yumn Suhaylah Yusoff

Faculty of Science and Technology,Universiti Sains Islam Malaysia, Bandar Baru Nilai, 71800 Nilai, Negeri Sembilan

Starting in November 2014, the Malaysia government has stopped the fuel subsidy and the fuel price is depending on the world crude oil price. However, the deregulation on the fuel price still has been a burden to all Malaysian as the cost of living increases. Hence, their income became insufficient to cover the monthly expenditures. This study undertakes to investigate the effects both directly and indirectly of fuel price fluctuations on household income in Malaysia due to subsidy removal. The total effects also being examined in this study The data from the Household Expenditure Survey (HES) conducted by Department of Statistics Malaysia is used in this study. Three different time periods are used which include year 2009/2010, 2014 and 2016. The households are differentiated into three income groups, namely Top 20% (T20), Middle 40% (M40) and Bottom 40% (B40). Then, the partial equilibrium approach is used to investigate the direct and indirect effect of fuel price fluctuations on three different income groups. This study found that group B40 experience the greatest significant indirect and total effects of fuel price fluctuations due to removal of subsidy. This is because this group spends the largest budget shares of their income on goods and services that affected by the increasing fuel price. For direct effect, there is not much different for all income groups after the subsidy removal. Then, the effects are found to be lower after the fuel subsidy removal for all ethnic groups. Thus, the findings could help the government in examining their policy of subsidy removal in order to be successful in helping the poor. Hence, this could improve the quality of life, especially for income groups B40 and M40.

Keywords:Fuel Price Fluctuation; Income Group; Partial Equilibrium; Subsidy removal

Paper ID : O021 A Hybrid of Quasi-Newton Method with CG Method for Unconstrained Optimization

N ’Aini¹, Mustafa Mamat², Mohd Rivaie³& Sulaiman Mohammed Ibrahim²

1Faculty of Computer and Mathematical Sciences, Universiti Teknologi MARA Cawangan Johor, Segamat, Malaysia

2Faculty of Informatics and Computing, Universiti Sultan Zainal Abidin, 21300, Terengganu, Malaysia

3Department of Computer Science and Mathematics, Universiti Teknologi MARA, Kuala Terengganu, Terengganu, Malaysia

The quasi-Newton is a well-known method for solving small to medium-scale unconstrained optimization prob-lems. This leads to many modifications to improve its performance, and one of them is by hybridizing it with another optimization method. In this study, the quasi-Newton method is combined with the ARM method, which is a type of conjugate gradient method. The resulting hybrid algorithm is globally convergent under exact line search.

Keywords:Quasi-Newton; Conjugate Gradient; Strong Wolfe Line Search; Unconstrained Optimization