Posters 2015

Posters presented at

(1) 8th Making Cities Liveable Conference, held at Pullman Melbourne on the Park from the 6 - 8 July 2015 

(2) The International Scientific Conference "Our Common Future under Climate Change", held at UNESCO, Paris Francis from 7-10 July 2015

 

Ken Doust / Charlotte Wang  

Title: Sustainability & Accessibility: The Next Step

 

 

 
   

Posters presented at the Civil and Environmental Student Poster Forum

19 February 2015

Nima Amini

Title: An Exploratory Assessment of the Associations between the Urban Environment and Obesity

Abstract:    Obesity and other chronic diseases are becoming more prevalent, as such researchers are trying to examine and combat this trend. A number of studies have shown an association between the built environment and walking. Other studies have shown an association between the built environment and health outcomes. However, contradictory associations have been reported. Review papers have assigned the blame to the well known Modifiable Aerial Unit Problem (MAUP). In this study a large set of data on urban environment in Sydney has been collated and processed to obtain a set of objective urban variables. These variables are guided by the principle concepts of how the urban environment may affect travel behaviour. The obesity and other individual data are obtained from the Household, Income and Labour Dynamics in Australia (HILDA) survey. The MAUP has been addressed through an innovative aggregation method, and have been shown to relate to the probability of being obese using a logit model. The self-selection issue has been addressed by removing the entries that selected their place of residence based on proximity to destinations.

 

David Arbis

Title: Modelling Strategic Interactions of Driver Manoeuvres

Abstract:  Game theory is the study of mathematical models of interaction. In this thesis, it is used to model the interaction of drivers on the road. The study contributes to safety design by better understanding behavioural norms of interaction.

Through a game-theoretic approach drivers are assumed to make the best action for themselves, taking into account the observed or predicted action of the competing individual. In this way drivers act according to the interdependence between their own actions and the actions of their competitor. Game theory thusly differentiates from traditional safety modelling approaches that consider drivers acting to one-directional influence. The approach is more holistic and results in a more precise analysis.

Further, as game-theoretic models investigate fundamental interaction behaviour, the study findings are generalizable beyond the traffic contexts studied.

Resources for this project include non-exhaustively the Australian School of Business Experimental Economics Laboratory (Quad1041) and the TRACSLab Driving Simulator (Building H20, LG8). 

 

  

Melissa Duell

Title: The impact of uncertainty on strategic network design projects

Abstract:    This work addresses the traffic network design problem (NDP), where the planner seeks the optimal locations in a network to which add vehicle capacity in order to minimize an objective, while staying within a budget. While traditionally deterministic, this work expands the NDP to include day-to-day uncertainties in travel demand and link capacity. Specifically, this project proposes a network design formulation that uses a strategic behaviour approach, in which total demand and link capacity are treated as random variables and a strategic user equilibrium results in fixed equilibrium link proportions.  The bi-level model is formulated, system performance metrics derived, and then a solution method is developed based on a tailored genetic algorithm. Results under varying levels of volatility reflect possible suboptimal project selection when using a deterministic modelling approach. 

 

Sisi Jian

Title: Hazard Based Modelling of Vehicle Selection in Carsharing Systems.

 Also presented at TRB 2015 (see below)

Tao Wen

Title: A maximum likelihood estimation of trip tables for the strategic user equilibrium (StrUE) Model

 Also presented at TRB 2015 (see below)

Kasun Wijayaratna

Title: Dynamic User Optimal traffic Assignment with Recourse

 Also presented at TRB 2015 (see below)

 

 

Posters presented at Transportation Research Board

Walter E. Washington Convention Center. Washington DC, USA
11-15 January 2015 

 

Melissa Duell

Title: The implications of volatility in day-to-day travel flow and road capacity on traffic network design projects'

Abstract: This work addresses the traffic network design problem when day-to-day uncertainties in travel demand and link capacity are taken into account. Specifically, this work proposes a network design formulation that uses a strategic behavior approach, where total demand and link capacity are treated as random variables and a strategic user equilibrium results in fixed equilibrium link proportions. The bilevel model is formulated, system performance metrics derived, and then a solution method is developed based on a tailored genetic algorithm. Results under varying levels of volatility reflect possible suboptimal project selection when using a deterministic modeling approach.

 

Milad Ghasrikhouzani

Title: A Novel System of Disaggregate Models for Travel Demand Modelling, Using Decision Tree and Random Forest Concepts

Abstract: This research attempts to address the gap between research and practical transport demand models. Evolution of travel demand models started from the early paper and pencil versions of conventional four-step models of the late 50s and proceeded to the activity-based models. During this transition the emphasis shifted from aggregate to disaggregate models, whereby researchers increasingly paid attention to individual decision making regarding daily activities. This alteration engendered a conspicuous disparity between aggregate and disaggregate models regarding their practicality, precision and policy sensitivity level. This disparity is the main attention of this paper. Unlike the four-step models, a highly disaggregate travel demand modelling structure is proposed in this paper in which trip purpose, mode of transport, time of day, commute distance and attributes of trip destination are modelled. Unlike activity-based models, the proposed method is computational efficient and cost effective making it attractive for small or medium size cities. Destination choice problem is decomposed from the framework by using the attributes of the destination rather than its actual location. All the decisions are modelled using Decision Tree (DT), a Modified version of Decision Tree (MDT), and Random Forest (RF). The model estimation practice is done using the Victorian Integrated Survey of Travel and Activity (VISTA) of 2007 as the training set, and its transferability is examined by utilizing the 2009 year of the same dataset. As the results show RF and DT system of models have a high disaggregate accuracy, whereas MDT and RF system of models predict the aggregate trip patterns in the test dataset more similar to the observations.

 

Milad Ghasrikhouzani

Title: A Residential Relocation Timing and Tenure Choice Modelling Structure: Competing Hazard Based or Cause-Specific Formulation?

Abstract: Land use is largely affected by people’s decisions on job and residential relocation as well as decisions by home developers and job providers. In regards to residential relocation decision-making, four fundamental components can be identified: 1) reason(s) of relocation, 2) relocation timing, 3) choosing the next living place, and 4) tenure choice. This paper focuses on the time of relocation and the tenure choice, using the hazard-based structure. A hazard-based model coupled with a logit model is compared against a competing hazard-based model. These two formulations represent two approaches that individuals may consider when evaluating their alternatives. A longitudinal data collected in Australia from the Household, Income, and Labour Dynamics Australia Survey (HILDA) is utilized in this study for developing the models. Several socio-demographic, economic and land use variables are employed to model the behaviour of decision-making unit which mainly consists of a household. It was found that the competing hazard model marginally outperforms the cause-specific structure.

 

Hanna Grzybowska

Title: A Decision Support System For A Real-Time Field Service Engineer Scheduling Problem With Emergencies And Collaborations

Abstract: We treat a dynamic routing and scheduling problem by repeatedly re-planning using a heuristic for solving a variety of the Vehicle Routing Problem. The problem we treat occurs in the situation when Field Service Engineers are assigned a sequence of jobs to attend. The jobs are geographically distributed and not all jobs to be undertaken are known in advance of planning. This dynamic occurrence of job requests is stochastic. Jobs are assigned an Emergency Level, which is the highest for the repair jobs involving a person in danger. In addition some jobs require two engineers to attend. We refer to such jobs as collaborative. Our approach re-schedules the pending jobs in an event-driven manner. The event-driven scheduling process ensures that jobs of high importance, with a high emergency level, are completed promptly. The proposed Decision Support System assists in the decision making concerning the management of Field Service Engineers, in the case when real-time information is available. Its architecture includes two main modules: Simulation Engine and Indigo Solver, using a flexible heuristic based on an Adaptive Large Neighborhood Search. We find that our approach of event-driven re-planning is able to plan for real-world scenarios using significantly fewer resources than are employed in practice.

Sisi Jian

Title: Hazard Based Modelling of Vehicle Selection in Carsharing Systems.

Abstract: Carsharing, as an alternative to private vehicle ownership, has spread worldwide in recent years due to its potential of reducing congestion, improving auto utilization rate and limiting the environmental impact of emissions release. To determine the most efficient allocation of resources within a carsharing programme, it is critical to understand what factors affect the users’ behaviour when selecting vehicles which have been little attended to in the literature. This study attempts to investigate the importance of users’ attributes and fleet characteristics on choice set formation behaviour in selecting vehicles using a Spatial Hazard Based Model (SHBM). In the SHBM model, “distance to the carsharing facility (vehicle pod)” is considered as the prospective decision criteria that carsharing users follow when selecting a vehicle among a set of alternatives This variable is analogous to the duration in a conventional hazard-based model Other factors, such as user socio-demographic attributes, vehicle characteristics, land use type of the trip origin, etc., collected from the Australian carsharing company GoGet are utilized for model estimation. A number of forms of parametric SHBMs are tested to determine the best fit to the data. The accelerated failure time model with a Log-logistic distribution was found to provide the best fit. The estimation results of the coefficients of the parameters suggest that users who do not own a private vehicle, frequent users, elderly users and users with a restriction in the ability to drive tend to consider closer vehicle pods. These findings can provide a starting point for carsharing organisations to optimize their pod locations and types of cars available at different pods to maximize usage. 

 

Mojtaba Maghrebi

Title: Machine Learning Fusion Based Technique for Predicting the Concrete Pouring Production Rate Based on Traffic and Supply Chain Parameters

Abstract: Being able to precisely predict the duration of construction tasks could provide insights for managers and enable them to handle their projects more efficiently. Most construction materials are supplied by out-sourced suppliers and are transferred via road transportation. Concrete is the most used construction material in the world and demand for concrete is ever increasing. In addition, a wide range of crews and machineries are involved in concrete based constructions tasks. So, being able to accurately estimate the concrete pouring task will potentially have both cost and time saving effects. Moreover, due to space limitations as well as technical obligations, fresh concrete is mixed at a Ready Mixed Concrete (RMC) depot and then hauled by trucks to constructions sites. Therefore, to predict the concrete pouring duration managers must consider both traffic and supply parameters. In this paper, a data structure is presented to cover these parameters and Machine Learner Fusion-Regression (MLF-R) is used to predict the production rate of concrete pouring tasks. A field database that covers a month of deliveries across a metropolitan area was gathered for evaluating the proposed method. The dataset includes over 2600 deliveries to 507 different locations. Finally, the MLF-R was tested with the proposed dataset and the results compared with ANN-Gaussian, ANN-Sigmoid and Adaboost.R2 (ANN-Gaussian) which are trained with the exact training sets. The results show that MLF-R obtained the least RMSE in comparison with other methods, and also acquired the least standard deviation of RMSE and correlation coefficient with the stability of this approach.

 

Mojtaba Maghrebi

Title: Optimality Gap of Experts’ Decisions in Concrete Delivery Dispatching

Abstract: Concrete delivery dispatching suffers from a lack of practical solutions and therefore, in the absence of automatic solutions, experts are hired to handle this task. In addition, the concrete delivery dispatching problem can be modelled mathematically but it can only solve up to medium sizes of this problem within a practical time. This paper attempts to answer the question of how much we can rely on experts' decisions. First, the concrete delivery problem is presented. Second, a benchmark for the problem is achieved; the heuristic approach is implemented for those instances where exact solutions are not computationally intractable. Finally, the experts' decisions are compared with the obtained benchmarks to assess the optimality gap of the experts. A field dataset which belongs to an active Ready Mixed Concrete (RMC) is used to evaluate the proposed idea. The results show that experts' decisions are near to optimum, with an average accuracy of 90%. However, after comparing individual decisions between optimisation models and the experts' decisions, we can conclude that optimisation models only try to achieve the lowest cost, while the expert prefers a more stable dispatching system at a little higher cost.

 

Emily Molyan

Title: Reliability- and Median-Based Identification of Toll Locations in Connected Vehicle Context

Abstract: In anticipation of pervasive onboard navigation and electronic payment associated with connected vehicle technology, universal dynamic tolling is a promising possibility for future congestion management. In order to take advantage of these advancements, planners must consider which metrics should be used for the selection and performance evaluation of tolls.

This work considers two metrics: median travel time and reliability measured with the 80th percentile travel time index. Each metric is tested for the identification of candidate tolling facilities and for the evaluation of the toll's performance. Using data mining techniques on measured volumes and speeds from California freeways, test tolls are simulated based on the two identification metrics

We find that both metrics contribute to the assessment of the toll's success, but travel time reliability is a stronger metric for determining the most effective locations for imposing tolls.

 

Tao Wen

Title: A maximum likelihood estimation of trip tables for the strategic user equilibrium (StrUE) Model

Abstract: This paper proposes a novel framework to estimate trip tables for the strategic user equilibrium traffic assignment model. The proposed framework uses a bi-level estimation model, where the upper-level is a new maximum likelihood estimation method and the lower-level is the strategic user equilibrium assignment model which accounts for some aspects of day-to-day volatility in traffic flow. The maximum likelihood method proposed in this paper illustrates its ability to utilize information from day-to-day observed link flows in order to provide a unique estimation of the total trip demand distribution. This is accomplished by passing the total trip demand distribution to the strategic user equilibrium model to produce a set of link flow distributions which can then be compared to the link level observations. The mathematical proof demonstrates the convexity of the model. In addition, a numerical analysis is conducted on a test network to illustrate the efficiency of the proposed framework.

 

Kasun Wijayaratna

Title: Dynamic User Optimal traffic Assignment with Recourse

Abstract: Limitations of static network equilibrium models have led to numerous research efforts in predicting the temporal and spatial traffic conditions throughout road networks.   Static models do not account for the dynamic nature of traffic.  Accordingly, new approaches to depict these scenarios have been formulated, such as Dynamic User Optimal (DUO).  This study extends DUO to DUO with recourse (DUOR) where a user can also alter their journey en-route dependent on the traffic conditions and the available information through ITS technology.
This study proposes the modelling framework Dynamic User Optimal with Recourse using a Cell Transmission Model (DUOR-CTM).  The model focuses on initially understanding whether a Dynamic User Optimal with Recourse (DUOR) solution arises and also determines the impact of information on the user optimal travel cost.  The study presents the results of a sample network and highlights the need to account for information in a dynamic context.