International Journal of Applied Operational Research

Service operations efficiency and competitive orientations in the telecommunication industry: a data envelopment analysis approach

A. J. Abiodun

Customers service centres have become a veritable tool for connecting with subscribers and projecting distinctive capabilities in competition in the telecommunication industry. Consequently, operational efficiency of these units is an essential variable in the design and adoption of cost leadership and differentiation strategies. This study is based on data obtained from twenty-five (25) customers service centres spread across four telecommunication service providers. Data envelopment analysis methodological approach was utilized to estimate efficiency of these service centres, types of scale and identify critical resources for improving performances. Results indicate an average technical efficiency scores range of 50.5%-80.2%, suggesting substantial waste of resources and weak capacity for pursuit of cost leadership strategy with scale efficiency scores range of 60.5%-86.4%. The study also found that 36% of these facilities are operating under increase returns to scale (IRS) regime, 40% (decreasing returns to scale, (DRS) and 24% under constant returns to scale regime. Furthermore, customers service personnel were identified as critical input variable to be given pre-emptive priority to change to enhance operations and capacity to strive for cost leadership, therefore, it is suggested that telecommunication firms invest more resources on technology to enhance capabilities of the service personnel for improved efficiency as well as for effective pursuit of cost leadership and differentiation strategies.

Offering a machine learning based algorithm, with the purpose of emergency brake during simulated driving based on EEG signal

M. Faridi Masouleh

Providing safe driving conditions has a great impact on reducing the amount of road accidents and deaths which are caused by them. The necessity of intelligent brake system for increasing the safety during driving has been taken into consideration in today’s cars. The automatic emergency braking system is responsible for informing the driver of impending accidents and using the ultimate potential of the vehicle's braking before a collision occurs. In this paper, the purpose is to predict the brake based on brain EEG signals. For this purpose, the standard bnci database which is defined in this field is used. The aim of the proposed method of this article, is to predict emergency brake during simulated driving, using after error propagation neural network algorithm. The innovative aspect of this paper is the combined use of the dimension reduction algorithm, after-error propagation neural network, and training by the means of K cross validation algorithm for reducing neural network learning error. The proposed method is trained with dataset feature vectors so that after feature vector entry, test recognizes that if emergency brake is necessary or not. Results obtained from proposed method show that the accuracy of this method is more than 90 percent, which has a better performance in comparison with other methods.

Fuzzy free replicability model with restricted variation

M. Jahani Sayyad Noveiri

In conventional data envelopment analysis (DEA) models, inputs, and outputs are usually considered as precise and continuous factors. Furthermore, inputs and outputs of inefficient decision-making units (DMUs) change arbitrarily for meeting the efficient frontier. Nevertheless, there are situations in the real world where the performance of DMUs with fuzzy and integer-valued measures must be evaluated while input and output variables are restricted by the decision-maker. Therefore, the current paper proposes a DEA-based method for assessing the relative efficiency of DMUs with imprecise and integer-valued factors when restricted variations are observed. To illustrate, the free replicability (FR) model is extended for incorporating fuzzy numbers and some visible limitations like restrictions on resources. Furthermore, the method is developed for situations where flexible measures are presented. A numerical example is used to illustrate the approach.

A Tri-objective model for multi-product multi-period inventory planning with substitutable goods and random demand

R. Sadeghian

The performed studies show that considering substitution goods can be profitable. The substitution is replacing one product with another in an inventory system. When a product has a shortage, a certain percentage of its demand can be replaced with similar goods. In this paper, a Tri-objective model with the substitution assumptions is considered. In here, the substitution means that there is a relation between different items that allows these items to be used instead of each other. Demands are considered probabilistic and there are some other assumptions as follows: planning is multi-period and multi product, inventory control parameters are fixed during the planning period. The shortage is allowed, but in the form of lost sale, the inventory of the beginning of the first period is very few (almost zero) and the remaining inventory at the end of each period will be moved to the next period. Objective functions are looking for maximizing the profit, minimizing the risk of facing slack and minimizing of the dissatisfaction arising from the substitution. Model is solved with two approaches: first with the LP-Metric method and next by two meta-heuristic algorithms such as NSGA-II and Deferential Evolution. Most researches have focused on profit maximization or costs minimization. The current paper considers a multi-product and multi-period triple-objective model. The goods may be substituted with similar ones. The results of solving the model indicate that if there is a relation between the products items, considering this relationship in modeling, will lead to improved results. A part of this improvement is a result of reduced maintenance cost. With the substitution of items, we can both increase our profits and sell items that their expiry date is near to finish (arrangement type of substitution) and avoid loses.

Sequential benchmarking to achieve the closest cross-sectional targets in DEA

B. Daneshian

The models that set the closest targets have made an important contribution to data envelopment analysis (DEA) as tool for best-practice benchmarking of decision making units (DMUs). These models may help define plans for improving the require less effort from the DMUs. One of the important issues in the process of benchmarking and target setting, is to set realistic and achievable targets for inefficient units. Because in practice and in the real world, we often face units that perform poorly and the targets for them are not available in one step, to solve this problem, in this study, an algorithm is presented that takes advantage of the onion layering method has three main advantages over other step-by-step benchmarking methods: firstly, in each step, it offers a better and closer target and benchmark to the manager sequentially. Secondly, by adjusting the number of jumps in the layers according to the conditions, it provides the possibility of more adjustments and flexibility in targets for the manager. Thirdly, by classifying the decision-making units based on the level of efficiency and performance, it specifies a benchmark and a realistic achievable target for the inefficient units at each stage. The proposed method has been implemented on the data of 24 Portuguese bank branches and And targets are specified sequentially for each ineffective unit.

Numerical solution for the fifth-order KdV equation by using spectral methods

M. Askaripour Lahiji

Nonlinear wave equations are more difficult to study mathematically and that no general analytical method for their solution exists. It is found that the Exponential Time Differencing (ETD) scheme requires the least steps to achieve a given accuracy, offers a speedy method at calculation time, and has exceptional stability properties in solving a nonlinear equation. This article explains how we applied the exponential integrators (ETDRK4) to semi-linear problems to solve the fifth-order KdV equation. To solve, we define a new integrating factor e^(-ik^5 t) and apply fast Fourier transform (FFT ) for spatial discretization. For this purpose, we solve the diagonal example of the fifth-order KdV equation via the exponential time differencing Runge-Kutta 4 method (ETDRK4). Implementation of the method is proposed by short Matlab programs.