State-space ARIMA for supply-chain forecasting

ARIMA is seldom used in supply chains in practice. There are several reasons, not the least of which is the small sample size of available data, which restricts the usage of the model. Keeping in mind this restriction, we discuss in this paper a state-space ARIMA model with a single source of error and show how it can be efficiently used in the supply-chain context, especially in cases when only two seasonal cycles of data are available. We propose a new order selection algorithm for the model and compare its performance with the conventional ARIMA on real data. We show that the proposed model performs well in terms of both accuracy and computational time in comparison with other ARIMA implementations, which makes it efficient in the supply-chain context.     

A novel intelligent approach for yaw position forecasting in wind energy systems

Yaw control systems orientate the rotor of a wind turbine into the wind direction, optimize the wind power generated by wind turbines and alleviate the mechanical stresses on a wind turbine. Regarding the advantages of yaw control systems, a k-nearest neighbor classifier (k-NN) has been developed in order to forecast the yaw position parameter at 10-min intervals in this study. Air temperature, atmosphere pressure, wind direction, wind speed, rotor speed and wind power parameters are used in 2, 3, 4, 5 and 6-dimensional input spaces. The forecasting model using Manhattan distance metric for k = 3 uncovered the most accurate performance for atmosphere pressure, wind direction, wind speed and rotor speed inputs. However, the forecasting model using Euclidean distance metric for k = 1 brought out the most inconsistent results for atmosphere pressure and wind speed inputs. As a result of multi-tupled analyses, many feasible inferences were achieved for yaw position control systems. In addition, the yaw position forecasting model developed was compared with the persistence model and it surpassed the persistence model significantly in terms of the improvement percent.     

Developing Cost Response Models for Company-Level Cost Flow Forecasting of Project-Based Corporations

The difficulty in applying the standard curve (S-curve) and cost-schedule integration (CSI) techniques for company-level cost flow forecasting in a project-based industry is the prerequisite of forecasting future unknown individual projects and contract classifications. By analyzing cost flows at the company level through a pool of macroeconomic and internal financial data, this paper proposes an innovative approach to firm-specific model estimation. First, a series of data transformations introduce linear relationships between cost, macroeconomic, and internal financial variables. Second, multivariate regression analysis is employed for initial model building. Third, for the purposes of model restructuring, a subsequent application of Yule–Walker estimates and incomplete principal component analysis is used. This paper uses a sample of four project-based construction firms to demonstrate model performance. Using this methodology, mean absolute percentage error (MAPE) values of the forecasting models range from 0.27 to 0.60%. As such, the transformed cost, macroeconomic, internal financial data could strongly predict company-level cost flow forecasting. While converting the predicted cumulative cost data to periodic cost flows, the MAPE values were augmented, ranging from 7.04 to 17.55%, thus, requiring future research.     

Forecasting Canadian nuclear power station construction costs

Because of the huge volume of capital required to construct a modern electric power generating station, investment decisions have to be made with as complete an understanding of the consequences of the decision as possible. This understanding must be provided by the evaluation of future situations. A key consideration in an evaluation is the financial component. This paper attempts to use an econometric method to forecast the construction costs escalation of a standard Canadian nuclear generating station (NGS). A brief review of the history of Canadian nuclear electric power is provided. The major components of the construction costs of a Canadian NGS are studied and summarized. A database is built and indexes are prepared. Based on these indexes, an econometric forecasting model is constructed using an apparently new econometric methodology of forecasting modelling. Forecasts for a period of 40 years are generated and applications (such as alternative scenario forecasts and range forecasts) to uncertainty assessment and/or decision-making are demonstrated. The indexes, the model, and the forecasts and their applications, to the best of the author's knowledge, are the first for Canadian NGS constructions.     

基于定性与定量分析的联络中心任务量预测法

任务量预测是联络中心人员排班的第一步,如何精确预测联络中心的任务量并相应地安排适当数量的座席人员来处理客户的需求是联络中心管理人员面临的首要问题.提出采用定性分析与定量分析相结合的任务量预测方法.对于给定的任务量数据,首先,采用观察法进行定性的分析区分异常和正常数据,然后,分别选定合适的定量预测方法进行预测.任务量定量预测方法包括趋势拟合、线性回归、移动平均、时间序列分解等预测方法.通过定性分析与定量分析相结合的综合应用方法,给出更准确的联络中心任务量预测方案,帮助管理人员最大程度的减少放弃和被拥堵在队列中的任务数量,设计制定可行性高的座席人员数量及人员排班.     

预测板式换热器污垢热阻的新方法

通过动态监测板式换热器冷却水污垢热阻及影响污垢热阻的松花江水水质参数（如pH值、溶解氧、铁离子、氯离子、细菌总数、浊度、电导率、化学需氧量、碱度和硬度等）变化。采用BP神经网络主成分分析、主成分回归、全要素BP神经网络三种预测方法建立板式换热器污垢热阻预测模型,选取1-15号样本为训练或回归拟合样本,16-20号样本为测试样本,并将三种方法的预测结果进行了对比。结果表明,三种方法均可对板式换热器污垢特性进行有效预测,而基于主成分分析的BP神经网络方法的预测结果误差小,优于另外两种方法。     

组合预测模型在电力系统负荷预测中的应用研究

本文基于灰色预测模型、滑动平均模型和指数平滑模型这三种单一预测模型,采用方差-协方差策略,建立组合预测模型。然后结合老挝电力系统的概况,对老挝的全国年用电量进行预测和分析。结果表明,组合预测模型的预测精度明显高于各单一预测模型,即组合预测模型的相对误差小于各单一预测模型的相对误差,说明组合预测模型具有相当的适用性和优越性。     

基于粒子群优化RBF神经网络原油含水率预测

原油含水率预测对于确定油井水、油层位以及估计原油产量有着非常重要意义。BP神经网络是最近常用的原油含水率预测方法,然而,由于BP神经网络存在容易陷入局部极小值、收敛速度慢等问题,影响了其预测的实用性和准确性,对此,提出基于粒子群优化RBF神经网络(PSO-RBFNN)的原油含水率预测方法,粒子群优化算法用于RBF神经网络参数优化。在分析原油含水率预测的影响因素基础上,建立粒子群优化RBF神经网络的原油含水率预测模型。实验结果表明,在原油含水率预测中,基于粒子群优化RBF神经网络比BP神经网络有着更高的预测精度。     

Modelling sustainable international tourism demand to the Brazilian Amazon

The Amazon rainforest is one of the world's greatest natural wonders and holds great importance and significance for the world's environmental balance. Around 60% of the Amazon rainforest is located in the Brazilian territory. The two biggest states of the Amazon region are Amazonas (the upper Amazon) and Pará (the lower Amazon), which together account for around 73% of the Brazilian Legal Amazon, and are the only states that are serviced by international airports in Brazil's north region. The purpose of this paper is to model and forecast sustainable international tourism demand for the states of Amazonas, Pará, and the aggregate of the two states. By sustainable tourism is meant a distinctive type of tourism that has relatively low environmental and cultural impacts. Economic progress brought about by illegal wood extraction and commercial agriculture has destroyed large areas of the Amazon rainforest. The sustainable tourism industry has the potential to contribute to the economic development of the Amazon region without destroying the rainforest. The paper presents unit root tests for monthly and annual data, estimates alternative time series models and conditional volatility models of the shocks to international tourist arrivals, and provides forecasts for 2006 and 2007.