A Comparative Analysis of Three Types of Tourism Demand Forecasting Models: Individual,Linear Combination and Non‐linear Combination

This paper investigates the combination of individual forecasting models and their roles in improving forecasting accuracy and proposes two non‐linear combination forecasting models using Radial Basis Function and Support Vector Regression neural networks. These two non‐linear combination models plus the standard Multi‐layer Perceptron neural network‐based non‐linear combination model are examined and compared with the linear combination models. The UK inbound tourism quarterly arrival data is used and the empirical results demonstrate that the proposed non‐linear combination models are robust and outperform the linear combination models that currently dominate in the tourism forecasting literature. Copyright © 2013 John Wiley & Sons, Ltd.     

旅游需求预测模型研究

建立基于月度数据的桂林漓江旅游航班、运量及游客的需求预测模型,运用指数平滑、SARIMA和Elman人工神经网络3种方法,并采用平均绝对误差(MAE)、均方百分比误差(RMSE)和平均绝对百分比误差(MAPE)评价模型预测效果。预测实例表明Elman人工神经网络模型更能反映时间序列的波动性,更适合桂林漓江旅游需求预测。     

Embedding economies of scale concepts for hub network design

We explore the idea of endogenous hub location on a network. In contrast to much of the literature, we propose that hub networks may emerge naturally out of a set of assumptions and conditions borrowed from equilibrium traffic assignment. To this end, we focus on applying a nonlinear cost function that rewards economies of scale on all network links. A model is presented and implemented in a GIS environment using both a 100-node intercity matrix and several synthesized interaction matrices. We compare solutions for different assumptions about network costs, and visualize the results. We find that under discounted conditions, network flow is re-routed to take advantage of the cost savings for amalgamation and that several cities emerge as centers through which large amounts of flow pass. Larger cities such as Los Angeles, New York and Chicago serve gateway functions. We also find that smaller cities such as Oklahoma City, Pittsburgh, Indianapolis, and Knoxville serve major gateway functions because of their locational advantages. Our paper should be of interest to the planner of a surface transportation system, or those interested in nodal concepts such as gateways and transport geography. Results are discussed in light of hub and spoke networks and suggestions are made for future research.     

Artificial neural network models for airport capacity prediction

This paper proposes artificial neural network models to predict the arrival/departure capacity of airports. Multilayer perceptron (MLP), recurrent neural networks (RNN), and long short-term memory (LSTM) models have been trained using capacity and meteorological data from Hartsfield–Jackson Atlanta International Airport (ATL) from 2013 to 2017. The models’ predictive performances were validated against the observed capacity of ATL in 2018. The qualitative and quantitative analysis of the trained models confirmed that the artificial neural networks approach is effective in predicting airport capacity. In addition, the transferability of the models for Boston Logan International Airport (BOS) is examined. Capacity prediction performance for BOS measures the transferability of the models trained with the ATL data. MLP showed good transferability without taking any other measures, and RNN and LSTM were able to predict the BOS capacity well after fine-tuning.     

An inventory-based simulation model for annual-to-daily temporal freight assignment

In the aggregate freight demand modeling literature, temporal assignment (annual to daily flows) is often oversimplified or neglected altogether. Unlike passenger flows, freight flows over the course of a year are not uniform and can vary significantly as the result of trade-offs between inventory and transportation cost management. We introduce the first temporal assignment model that explicitly considers these trade-offs for aggregate freight forecasting. A two-stage model is proposed that first decomposes aggregate annual zonal flows to firm group annual flows using a supply chain network model, which are then temporally assigned by simulating purchase order transactions throughout supply chains. Lot sizes are estimated with an Economic Order Quantity (EOQ) model and calibrated with monthly inventory data. The result is an aggregate-disaggregate-aggregate model that fits into aggregate freight forecasting models but makes use of more disaggregate logistical data. The model is illustrated with a simple replicable example, followed by a case study conducted with California statewide data to break out the distributed zonal flows into average daily volumes for network assignment. Calibration results using 2007 IMPLAN data showed a median percentage difference of simulated annual flows from FAF³ data of 2.38%, and a median percentage difference of simulated inventories from IMPLAN data of 4.85%, which suggests an excellent fit. Empirical validation results showed the model outperforms fixed factor approaches in mean value accuracy by 15–31%.     

Temporal aggregation and spatio-temporal traffic modeling

Traffic forecasting is crucial for policy making in the transport sector. Recently, Selby and Kockelman (2013) have proposed spatial interpolation techniques as suitable tools to forecast traffic at different locations. In this paper, we argue that an eventual source of uncertainty over those forecasts derives from temporal aggregation. However, we prove that the spatio-temporal correlation function is robust to temporal aggregations schemes when the covariance of traffic in different locations is separable in space and time. We prove empirically this result by conducting an extensive simulation study on the spatial structure of the Milan road network.     

Stochastic fleet deployment models for public bicycle rental systems

This paper presents two stochastic bike deployment (SBD) models that determine the optimal number of bicycles allocated to each station in a leisure-oriented public bicycle rental system with stochastic demands. The SBD models represent the stochastic demands using a set of scenarios with given probabilities. A multilayer bike-flow time-space network is constructed for developing the models, where each layer corresponds to a given demand scenario and effectively describes bicycle flows in the spatial and temporal dimensions. As a result, the models are formulated as the integer multi-commodity network flow problem, which is characterized as NP-hard. We propose a heuristic to efficiently obtain good quality solutions for large-size model instances. Test instances are generated using real data from a bicycle rental system in Taiwan to evaluate the performance of the models and the solution algorithm. The test results show that the models can help the system operator of a public bicycle system make effective fleet deployment decisions.     

Forecasting inland waterway grain traffic

The purpose of this paper is to address a neglected area of water transportation forecasting – short-term forecasting of inland waterway traffic. A time series model is used to forecast Mississippi River Lock 27 grain tonnage for the 1989:1–1999:4 period. The model was selected on the basis of several measures of goodness of fit and out-of-sample forecasting performance. The out-of-sample forecasting performance of the model was good, as the percentage difference between the year 2000 actual and forecasted tonnage was less than 5.5% for three of the four quarters and only about 2% for the year.     

Performance measurement of a transportation network with a downtown space reservation system: A network-DEA approach

We propose an evaluation approach for a novel travel demand management strategy known as the downtown space reservation system (DSRS). This approach takes into account three perspectives, i.e., transportation service provider’s, the user’s, and the community’s and is based on network-Data Envelopment Analysis (DEA) where the perspectives are inter-related through intermediate inputs/outputs. Two types of network-DEA models (radial and slacks-based models) are considered. An example is provided using data propagated from a microscopic traffic simulation model of the DSRS. The results show that individual node performance can drive network DEA performance and that this information can inform future designs of the DSRS.     

Varying Elasticities and Forecasting Performance

This study assumes that tourists' demand reactions to income and price changes are asymmetric at different phases of the business cycle. In order to test this hypothesis, we analyzed the demand for international tourism in five source markets using a modified growth rate (MGR) model. The empirical evidence demonstrates that income elasticity is indeed asymmetric across the business cycle in four source markets. In addition, asymmetric price effects were found for one source market. To compare forecasting performance, we also estimated a time‐varying parameter (TVP) model. The results show that the MGR model generally outperforms the TVP model. Copyright © 2013 John Wiley & Sons, Ltd.     

Levels of internationalization in the container shipping industry: an assessment of the port networks of the large container shipping companies

This paper examines the internationalization level of the world’s largest container shipping companies by examining their port networks. The results of our analysis show that only three of the nineteen companies could be considered as being truly ‘International’ and service the ‘lion’s share’ of the major and minor ports distributed across the globe. The port networks of the other companies in our survey, whilst having individualized features, shared common characteristics. Guided by the ‘Uppsala Model of Incremental Internationalization’ (Johanson and Vahlne, 1977), we were able to identify four internationalization levels of a container shipping company that are indicated by their port network.We argue that the port network plays an additional role as it contributes to the internationalization level of the container shipping companies. This is important because the internationalization level affects the container shipping company’s ability to expand its customer base both at a local and international level. Local customers would have access to a larger international destination network and international customers would be able access a larger local distribution network without transhipping between different companies.Amongst other issues, we were able to determine that, for the analyzed container shipping companies, ‘Internationalization’ of a port network means including or excluding ports in specific regions or sub-regions in their own network, and that a company’s home base or history plays a significant role in this development.     

Modelling growth principles of metropolitan public transport networks

The development of metropolitan public transport networks often involves choosing between investing in extending radial lines or constructing ring connections. While the former enlarges network coverage the latter enhances network connectivity and reduces the need to perform detours. Moreover, investments might be better directed at increasing the capacity of already existing infrastructure. In this study we address the following question: how do transport networks in metropolitan areas evolve over time and how can we effectively model this growth as function of demand and cost function? The goal of this study is to determine the fundamental relations between population distribution, modal costs on the prevailing network structure and its evolution. The approach taken in this study offers a theoretical contribution to the field of transport network growth by combining principles from several research streams: transport geography, economics of network growth and network science. We propose an iterative investment model network analysis framework. The results of the network growth experiments manifest an overall trend in network growth with an early phase of expansion of the network, followed by a period of intensification manifested in capacity increments and finally adding some links that contribute to its densification. Furthermore, our findings suggest that bus networks include more ring-radial connections than Light Rail Train and Metro networks which are more concentrated on radial connections.     

Taiwan's international tourism: a time series analysis with calendar effects and joint outlier adjustments

In this paper, we examine monthly tourist arrivals from Japan, Hong Kong and the USA between January 1971 and December 2008. Our purpose is to find events or variables that affect Taiwan's international tourism. We find that the Chinese New Year has a positive effect on tourist arrivals from Hong Kong, but negative effects for other countries. Through outlier detection, we obtain a better understanding of the effects of non‐recurring events that have impacted Taiwan's international tourism. Using transfer function model with automatic outlier detection and adjustment, we find that the exchange rate influences tourist arrivals from Japan and Hong Kong. Copyright © 2010 John Wiley & Sons, Ltd.     

Stakeholder Network Integrated Analysis: The Specific Case of Rural Tourism in the Portuguese Peneda‐Gerês National Park

This study addresses the structure, nature and composition of stakeholder networks related to rural tourism in a Portuguese National Park (Peneda‐Gerês National Park). In particular, it examines their interactions along the following dimensions: marketing information (marketing exchanges such as flyers, tourism information and information on specific events), administrative resources (logistic and technical support), human resources (allocation of human resources), training (training of local stakeholders in the area of promotion) and financial resources (financial support). In terms of methodology, it uses both qualitative and quantitative techniques. The analysis of the type and nature of interactions among different stakeholders within a specific rural tourism destination is the main focus of this study. Among these, the marketing information network shows both a high number of connections and high centrality indices, which may indicate that there is a significant share of information about common projects, products and services among stakeholders. In contrast, the training network evidenced the opposite behavior pattern. The study also contributes to a better understanding of how different local stakeholders are positioned in a complex structure of interactions that are critical in providing useful directions to maximize potential opportunities that may contribute to promote rural tourism and local development efficiently. Several limitations and implications are offered in the final section. Copyright © 2014 John Wiley & Sons, Ltd.     

Tourist satisfaction and beyond: tourist migrants in Mallorca

This exploratory study analyses the role of tourist experience, especially tourist satisfaction, in subsequent migration from the UK to the municipality of Calviá, Mallorca, Spain. An overview and commentary is provided of theory relating to both tourist satisfaction and migration. Through a series of elite interviews, in‐depth micro‐studies and a questionnaire, which provides a small quantitative component, it is established that the motivation given by tourist satisfaction can be recognised as a key catalyst for migration. Classic migration motivations cannot be entirely discounted, and the migration decisions are multi‐stage through time. But tourist satisfaction—and particularly the two satisfaction components of performance and emotion—is especially relevant in the micro‐perspective of migration decision‐making. Copyright © 2008 John Wiley & Sons, Ltd.     

Persistence Characteristics of Tourism Arrivals to Australia

This paper examines the degree of persistence in international monthly arrivals to Australia by using data disaggregated by tourism‐source countries. We employ two competing models, which are very general in the sense that they include (seasonal and non‐seasonal) unit roots as particular cases of interest. The first model is based on a long‐memory process in the non‐seasonal part of the series along with a short‐memory autoregressive (AR) seasonal structure. The second model is based on a long‐memory process for the seasonal structure of the series, the short‐term evolution being described through a non‐seasonal AR(1) process. Results based on the residuals and forecasting assessment indicate that the second model is preferable in terms of fitting the data. We provide persistence ranking of all countries included in the study and discuss the managerial implications of the main findings. Copyright © 2011 John Wiley & Sons, Ltd.     

Managing lumpy demand for aircraft spare parts

This paper deals with effective forecasting methods for typically lumpy demand for aircraft spare parts, and analyzes the behavior of forecasting techniques when dealing with lumpy demand. Twenty forecasting techniques are considered and tested and historical data from Alitalia are used to analyze and compare their performance. The results demonstrate that item lumpiness is the dominant parameter and show that demand forecasting for lumpy items is a complex problem; results from previous studies are not very accurate. The best approaches are found to be weighted moving averages, the Croston method, and exponentially weighted moving average models.     

Mitigating Gender Inequality in Rural Regions: The Effects of Tourism Employment in Namibia

We examine the effects of tourism employment on gender inequality in Namibia. We find that tourism contributes to local livelihoods by providing opportunities for cash income through employment, craft sales and shareholder profits from tourism enterprises. In general, male‐headed households are economically better off. Female‐headed households with tourism employment are better off than those without and are as equally well‐off as male‐headed households. We find no significant statistical difference between male‐headed households with tourism employment and those without it. These findings suggest that tourism provides specific advantages for women and helps mitigate gender inequality in areas where tourism employment is available. Copyright © 2014 John Wiley & Sons, Ltd.     

Attracting Convention and Exhibition Attendance to Complex MICE Venues: Emerging Data from Macao

This study presents an importance‐performance analysis of multi‐level attributes (event, facility and destination) evaluated by delegates attending an exhibition event in a ‘complex meetings, incentive, convention or exhibition (MICE) venue’ in greater China (mainland China, Hong Kong, Macao and Taiwan). The study's findings expound the relevance of various attributes in light of the emergence of complex MICE venues and destination resorts and, in particular, emphasizes the relative importance of destination – vis‐à‐vis facility – and core event‐related attributes towards determining exhibition attendance. Copyright © 2012 John Wiley & Sons, Ltd.     

Ecotourism impacts in the Nicoya Peninsula,Costa Rica

Ecotourism promotes responsible travel to natural areas, environmental conservation and the well‐being of local communities. Eco‐lodges are an important component of ecotourism ventures but an infrequently researched component of this field. Considering their influence on the natural environment (design and operation) and local communities (employment practices and purchases), the success of ecotourism depends, in part, on the performance of eco‐lodges. This project studies the effects of the Punta Islita (PI) eco‐lodge on the Nicoya Peninsula, Costa Rica. An interdisciplinary nested‐scale analysis, combining guest and household interviews with multi‐temporal remote sensing analysis of forest cover change of the lodge and surrounding areas, is used to evaluate the environmental, economic and social impacts of ecotourism in the region. Our results show significant positive contributions of the PI eco‐lodge on forest cover, environmental conservation, and local economic incomes within the surrounding communities. For local livelihoods, the PI eco‐lodge was seen as having positive social, cultural and economic impacts for nearly all societal variables for both employees and their neighbours. The PI eco‐lodge was also perceived as reducing alcoholism, drug addiction and prostitution, where conventional tourism on the Peninsula was shown to increase these ‘societal ills’. Land value and product pricing were the few variables believed to have increased as a result of tourism on the Peninsula. For conservation, the PI eco‐lodge property had the highest rates of reforestation within the Nicoya Peninsula and remains the scale most reforested in both forest cover change and total forest cover. In fact, at the landscape scale, we find that the Pacific coast of the Nicoya, where the bulk of ecotourism occurs, has undergone reforestation, whereas forest interiors have been deforested. Historically, reforestation occurred as cattle ranching credit programs were halted by the government and households in the area left to find better job opportunities. The PI eco‐lodge, as a source of good employment, resulted in worker migration back to the surrounding area, resulting, in some cases, in increased deforestation. Overall, we feel that the PI eco‐lodge serves as an example of successful ecotourism. However, increasing development in the region, in particular by standard hotel operations and large condo developments, seeks to capitalise on the region's natural beauty and may reverse land cover trends if they are not accompanied by adequate forest conservation strategies and stresses the importance of monitoring and assessing the impacts of accommodations tied to nature‐based tourism operations. Copyright © 2010 John Wiley & Sons, Ltd.