A design of experiment based procedure for real-time forecasting

  • Pennsylvania State University Harrisburg
To read the full-text of this research, you can request a copy directly from the authors.


Time series methods base their forecasts on extrapolations from past patterns and inter-relationships. Consequently, they work well only when the future is similar to the past or when changes (by chance) happen to cancel out; they are also quite handicapped when it comes to the consideration of environmental factors. In a turbulent environment with high uncertainty, the need for accurate forecasts is paramount. Hence, this paper proposes a design-approach, which incorporates the principles of Design of Experiments (DOE) into the real-time forecasting model, such as causal methods, to minimise the standard error. DOE is employed to select the most significant factors while forming the causal equation and further to perfect the coefficients of these factors. A pilot study data has been used to validate the proposed model.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... In most of the mentioned methods, it is claimed that, due to high fluctuations of energy consumption in high energy consumption industries, conventional regression models do not forecast energy consumption correctly and precisely. As Vinay Arul et al. (2007) mentioned in an uncertain environment, the need for accurate forecast is undeniable. They proposed a design of experiment-based procedure into real time forecasting model and rejected time series method as a suitable approach for real time forecasting. ...
Gas consumption plays a vital role in socio-economic development of most countries. This study presents a fuzzy regression algorithm for forecasting gas consumption based on standard economic indicators. The proposed approach utilises the most standard independent variables for the regression models. The standard input indicators are gross domestic production, population, number of vehicles and actual price of gasoline. To show the applicability and superiority of the proposed fuzzy regression model, the data for gasoline consumption in Iran from 1992 to 2005 is used. The results show that the fuzzy regression approach provides better solution than conventional regression for gas consumption estimation in Iran. The algorithm may be used by policy makers to accurately foresee the behaviour of gas consumption in various regions.
... Results indicate the good performance of CLPSO in solving multi-modal problems compared with earlier versions. Arul et al. (2007) proposed a design approach for the real time forecasting with the aid of design of experiments. See and Wong (2008) formulated the facility layout problem (FLP) as quadratic assignment problem using exact and heuristic techniques. ...
Full-text available
This paper presents a modification to the particle swarm optimisation (PSO) to tackle two difficulties observed in many applications: premature convergence of the solution, and the degree of confidence of the decision maker. This approach, known as triggered particle swarm optimisation, treats the problem in a dynamic environment and making each particle reset its record of best positions. This approach treated the PSO by triggering the particle swarm optimiser in a dynamic environment, making each particle reset its record of its best position. This, in turn, avoids making position and velocity changes based on outdated information. Due to random-based nature, a statistical confidence interval estimation approach is developed around the returned optimum at different levels. The proposed algorithm, triggered particle swarm optimisation (T-PSO), performs significantly better than the original PSO and the new particle swarm optimisation (NPSO) discussed in references.
... 06) compared the empirical performance of various forecasting models in assessing the effects of policy variables, legal changes, and traffic security campaigns. They also compared aggregated versus disaggregated forecasts of the main accident variables in order to examine the robustness of the forecasting improvement from using disaggregated data. Arul et al. (2007) proposed a design-approach, which incorporates the principles of Design of Experiments into the real-time forecasting model, such as causal methods, to minimise the standard error. Franses and Dijk (2005) examined the forecasting performance of various models for seasonality and non-linearity for quarterly industrial production series o ...
Full-text available
This paper develops an artificial neural network (ANN) model to forecast the optimum demand as a function of time of the year, festival period, promotional programmes, holidays, number of advertisements, cost of advertisements, number of workers and availability. The model selects a feed-forward back-propagation ANN with 13 hidden neurons in one hidden layer as the optimum network. The model is validated with a furniture product data of a renowned furniture company. The model has also been compared with a statistical linear model named Brown’s double smoothing model which is normally used by the furniture companies. It is observed that ANN model performs much better than the linear model. Overall, the proposed model can be applied for forecasting optimum demand level of furniture products in any furniture company within a competitive business environment.
The support of containerization to trade development demands an efficient solution method for the container loading problem in order to reduce shipment and handling time. Hence, the stowage planning of containers is critical to provide speedy delivery of resources from the area of supply to the area of demand. Moreover, information on container terminal activities, structure of ship, and characteristics of containers is distributed among stowage planners. This information imposes constraints, and so the master bay plan problem (MBPP) becomes NP-hard. Therefore, a multi-agent systems (MAS) methodology is designed to effectively communicate the information and solve the MBPP sustainably. In the designed MAS methodology, an information exchange system (IES) is created for stowage planners to bid for ship slots in each experimental iterative combinatorial auction (ICA) market. The winner in the ICA experiments is provided with the ship slots, and the entire bay plan is prepared. Further, the ship-turnaround time is validated using the data obtained from the benchmark problem of Wilson and Roach (1999).
ResearchGate has not been able to resolve any references for this publication.