Search results

The purpose of this paper is to provide an attempt to evaluate the risk-adjusted performance of international mutual funds using the risk statistic generated by the mean absolute deviation (MAD) and promote the ability of portfolio managers and investors to make the logical decisions for selecting different funds using the new optimized measures.

This study evaluates the performance of 50 international mutual funds using optimized risk-adjusted measures by the MAD over the monthly period 2001-2010. Using 50 linear programming models, the MAD is first computed by the linear programming models, and then seven performance measures of Treynor, Sharpe, Jensen’s α, M2, information ratio (IR), MSR, and FPI are optimized and proposed by the MAD to evaluate the mutual funds.

The empirical evidence detects that the MAD is an important determinant to evaluate the funds’ performance. Using the MAD statistic, this paper shows that new optimized measures are mostly over-performed by the benchmark index; in addition, these optimized measures have close correlation with each other. The results, therefore, detect the importance of using new optimized measures in evaluating the mutual funds’ performance.

The result of this study can be directly used as an initial data for decision of investors and portfolio managers who are seeking the possibility of participating in the global stock market by the international mutual funds.

This paper is the first study which optimizes the variance of returns in the MAD framework for each fund to propose new seven optimized measures of Treynor, Sharpe, Jensen’s α, M2, IR, MSR, and FPI.

Numerous articles have been written to prove or to disapprove the hypothesis of market efficiency. The purpose of this paper is to apply the forecast accuracy measure, mean absolute deviation (MAD), to check the validity of the hypothesis.

Forecast accuracies from applying different simple moving average methods to independently identically distributed (i.i.d.) or near i.i.d. normal time series are assessed by MAD. When moving period n is greater than m, then the mean of the MADs from the MA with n moving periods will be smaller than the mean of the MADs from the MA with m moving periods.

In this study, when different MAs are applied to four near i.i.d. finance time series from Fama’s papers, the MAD cannot distinguish the differences among MA methods with various moving periods. This contradiction implies that the four finance time series in Fama’s papers may not be i.i.d and implies that the market is not efficient.

The finding is only based on simulation and four near i.i.d. time series studied in Fama’s papers in 1965 and 1970.

The study shows that that the differences of the rates of returns from Johns Manville, Goodyear, Owens Illinois, and General Electric studied are not i.i.d. and that the market is not efficient. It refutes what Fama (1965, 1970) has claimed.

When the market is not efficient, investors may gain profit from the market.

Based on the literature review, this is the first study to use the forecast accuracy measure, MAD, for market efficiency.

This is a study of forecasting models that aggregate monthly times series into bimonthly and quarterly models using the 1,428 seasonal monthly series of the M3 competition of Makridakis and Hibon (2000). These aggregating models are used to answer the question of whether aggregation models of monthly time series significantly improve forecast accuracy. Through aggregation, the forecast mean absolute deviations (MADs) and mean absolute percent errors (MAPEs) were found to be statistically significantly lower at a 0.001 level of significance. In addition, the ratio of the forecast MAD to the best forecast model MAD was reduced from 1.066 to 1.0584. While those appear to be modest improvements, a reduction in the MAD affects a forecasting horizon of 18 months for 1,428 time series, thus the absolute deviations of 25,704 forecasts (i.e., 18*1,428 series) were reduced. Similar improvements were found for the symmetric MAPE.

The optimization of investment portfolios is a topic of major importance in financial decision making, with many relevant models available in the relevant literature. The purpose of this paper is to perform a thorough comparative assessment of different bi-objective models as well as multi-objective one, in terms of the performance and robustness of the whole set of Pareto optimal portfolios.

In this study, three bi-objective models are considered (mean-variance (MV), mean absolute deviation, conditional value-at-risk (CVaR)), as well as a multi-objective model. An extensive comparison is performed using data from the Standard and Poor’s 500 index, over the period 2005–2016, through a rolling-window testing scheme. The results are analyzed using novel performance indicators representing the deviations between historical (estimated) efficient frontiers, actual out-of-sample efficient frontiers and realized out-of-sample portfolio results.

The obtained results indicate that the well-known MV model provides quite robust results compared to other bi-objective optimization models. On the other hand, the CVaR model appears to be the least robust model. The multi-objective approach offers results which are well balanced and quite competitive against simpler bi-objective models, in terms of out-of-sample performance.

This is the first comparative study of portfolio optimization models that examines the performance of the whole set of efficient portfolios, proposing analytical ways to assess their stability and robustness over time. Moreover, an extensive out-of-sample testing of a multi-objective portfolio optimization model is performed, through a rolling-window scheme, in contrast static results in prior works. The insights derived from the obtained results could be used to design improved and more robust portfolio optimization models, focusing on a multi-objective setting.

The moving median method is suggested as an alternative for short‐term forecasting under some of the standard normal; Student t with degrees of freedom 1, 2 and 3; Cauchy; Chi‐square with degrees of freedom 1 and 2; exponential; and standard normal with outliers. When the Mean Absolute Deviation (MAD) and the Mean Square Error (MSE) are used as the criteria for evaluating forecasting accuracy, the moving average technique is superior to the moving median technique only for time series simulated from the standard normal distribution. The moving median technique is superior to the moving average technique for the fat‐tailed distributions; for example, t‐distributions of degrees 1, 2 and 3, Cauchy distribution, and the contaminated normal cases. An example shows the moving median technique responds to the level changes faster than the moving average technique. An illustrative example is also given for a practical data set.

Scheduling of aircraft maintenance operations is a gap in the literature. Maintenance times should be determined close to the real-life to schedule aircraft maintenance operations effectively. The learning effect, which has been studied extensively in the machine scheduling literature, has not been investigated on aircraft maintenance times. In the literature, the production times under the learning effect have been examined in numerous studies but for merely manufacturing and assembly lines. A model for determining base and line maintenance times in civil aviation under the learning effect has not been proposed yet. It is pretty challenging to determine aircraft maintenance times due to the various aircraft configurations, extended maintenance periods, different worker shifts and workers with diverse experience and education levels. The purpose of this study is to determine accurate aircraft maintenance times rigorously with a new model which includes the group learning effect with the multi-products and shifts, plateau effect, multi sub-operations and labour firings/rotations.

Aircraft maintenance operations are carried out in shifts. Each maintenance operation consists of many sub-operations that are performed by groups of workers. Thus, various models, e.g. learning curve for maintenance line (MLC), MLC with plateau factor (MPLC), MLC with group factor (MGLC) were developed and used in this study. The performance and efficiency of the models were compared with the current models in the literature, such as the Yelle Learning model (Yelle), single learning curve (SLC) model and SLC with plateau factor model (SLC-P). Estimations of all these models were compared with actual aircraft maintenance times in terms of mean absolute deviation (MAD), mean absolute percentage error (MAPE) and mean square of the error (MSE) values. Seven years (2014–2020) maintenance data of one of the top ten maintenance companies in civil aviation were analysed for the application and comparison of learning curve models.

The best estimations in terms of MAD, MAPE and MSE values are, respectively, gathered by MGLC, SLC-P, MPLC, MLC, SLC and YELLE models. This study revealed that the models (MGLC, SLC-P, MPLC), including the plateau factor, are more efficient in estimating accurate aircraft maintenance times. Furthermore, MGLC always made the closest estimations to the actual aircraft maintenance times. The results show that the MGLC model is more accurate than all of the other models for all sub-operations. The MGLC model is promising for the aviation industry in determining aircraft maintenance times under the learning effect.

In this study, learning curve models, considering groups of workers working in shifts, have been developed and employed for the first time for estimating more realistic maintenance times in aircraft maintenance. To the best of the authors’ knowledge, the effect of group learning on maintenance times in aircraft maintenance operations has not been studied. The novelty of the models are their applicability for groups of workers with different education and experience levels working in the same shift where they can learn in accordance with their proportion of contribution to the work and learning continues throughout shifts. The validity of the proposed models has been proved by comparing actual aircraft maintenance data. In practice, the MGLC model could efficiently be used for aircraft maintenance planning, certifying staff performance evaluations and maintenance trainings. Moreover, aircraft maintenance activities can be scheduled under the learning effect and a more realistic maintenance plan could be gathered in that way.

Organisations either keep spares for their own use, or‐for‐sale to other organisations. In either case, the ultimate need is to be able to replace worn or defective parts in operational machinery or equipment. In an economic sense, spares are kept to meet the needs of the situation in the cheapest way.

To propose an improved algorithm for the state estimation of distribution networks based on the unscented Kalman filter (IUKF). The performance comparison of unscented Kalman filter (UKF) and newly developed algorithm, termed Improved unscented Kalman Filter (IUKF) for IEEE-30, 33 and 69-bus radial distribution networks for load variations and bad data for two measurement noise scenarios, i.e. 30 and 50 per cent are shown.

State estimation (SE) plays an instrumental role in realizing smart grid features like distribution automation (DA), enhanced distribution generation (DG) penetration and demand response (DR). Implementation of DA requires robust, accurate and computationally efficient dynamic SE techniques that can capture the fast changing dynamics of distribution systems more effectively. In this paper, the UKF is improved by changing the way the state covariance matrix is calculated, to enhance its robustness and accuracy under noisy measurement conditions. UKF and proposed IUKF are compared under the cummulative effect of load variations and bad data based on various statistical metrics such as Maximum Absolute Deviation (MAD), Maximum Absolute Per cent Error (MAPE), Root Mean Square Error (RMSE) and Overall Performance Index (J) for three radial distribution networks. All the simulations are performed in MATLAB 2014b environment running on an hp core i5 laptop with 4GB memory and 2.6 GHz processor.

An Improved Unscented Kalman Filter Algorithm (IUKF) is developed for distribution network state estimation. The developed IUKF is used to predict network states (voltage magnitude and angle at all buses) and measurements (source voltage magnitude, line power flows and bus injections) in the presence of load variations and bad data. The statistical performance of the coventional UKF and the proposed IUKF is carried out for a variety of simulation scenarios for IEEE-30, 33 and 69 bus radial distribution systems. The IUKF demonstrated superiority in terms of: RMSE; MAD; MAPE; and overall performance index J for two measurement noise scenarios (30 and 50 per cent). Moreover, it is shown that for a measurement noise of 50 per cent and above, UKF fails while IUKF performs.

UKF shows degraded performance under high measurement noise and fails in some cases. The proposed IUKF is shown to outperform the UKF in all the simulated scenarios. Moreover, this work is novel and has justified improvement in the robustness of the conventional UKF algorithm.

Forecasting is a vital part of the planning process of most private and public organizations. A number of extant measures: Mean Absolute Deviation (MAD), Mean Square Error (MSE) and Mean Absolute Percentage Error (MAPE), have been used to assist in judging the forecast accuracy, and concomitantly, the consequences of those forecasts. In this paper we introduce, evolve, and implement a practical and effective method for assessing the accuracy of forecasts, the Percent Forecast Error (PFE). We test and evaluate the PFE, and modified optimized PFE (MOPFE), against the MAD, MSE, and MAPE measures of forecast accuracy using three time series datasets.