Do Safe Bike Lanes Really Slow Down Cars? A Simulation-Based Approach to Investigate the Effect of Retrofitting Safe Cycling Lanes on Vehicular Traffic.

Cycling is a sustainable transportation mode that provides many health, economic and environmental benefits to society. Cities with high rates of cycling are better placed to address modern challenges of densification, carbon-neutral and connected 20-min neighbourhood goals. Despite the known benefits of cycling, participation rates in Australian cities are critically low and declining. Frequently, this low participation rate is attributed to the dangers of Australian cycle infrastructure that often necessitates the mixing of cyclists with car traffic. In addition, residents of car-dependent Australian suburbs can be resistant to the installation of cycle infrastructure where threats to traffic flow, or decreased on-street parking availability are perceived and the prohibitive cost of reconfiguration of other infrastructure maintained by the local councils to retrofit safe bike paths. This study investigates the effects on traffic behaviour of retrofitting safe, separate cycling lanes into existing residential streets in a Melbourne suburb suitable for accessing the primary neighbourhood destinations. We utilise only the widths available on the existing roadway of these streets, with minimal incursion on other facilities, such as the vehicle network and parking. Using only the existing roadway reflects the common need for municipal asset managers to minimise disruption and costs associated with street redesign. Using a traffic simulation approach, we modelled travel demand that suits suburban trips to services and shops, and we selectively applied separate cycling lanes to suitable residential streets and varied the effect of lowering speed limits. Simulations show that the selective inclusion of safe cycling lanes in some streets leads to a mere 7% increase in the average car travel times in the worst case, while requiring cyclists to increase their travel distance only marginally to avoid streets without dedicated cycling lanes. These results demonstrate that reasonable compromises are possible to make suburbs safer for cyclists and bring them closer to the 20-min neighbourhood goal. There is significant potential to enhance the result by including more street types and alternative designs. The results can inform councils in their cycle path infrastructure decisions and disprove assumptions about the influence of cyclists on car infrastructure.

Cognitive Service Virtualisation: A New Machine Learning-Based Virtualisation to Generate Numeric Values.

Continuous delivery has gained increased popularity in industry as a development approach to develop, test, and deploy enhancements to software components in short development cycles. In order for continuous delivery to be effectively adopted, the services that a component depends upon must be readily available to software engineers in order to systematically apply quality assurance techniques. However, this may not always be possible as (i) these requisite services may have limited access and (ii) defects that are introduced in a component under development may cause ripple effects in real deployment environments. Service virtualisation (SV) has been introduced as an approach to address these challenges, but existing approaches to SV still fall short of delivering the required accuracy and/or ease-of-use to virtualise services for adoption in continuous delivery. In this work, we propose a novel machine learning based approach to predict numeric fields in virtualised responses, extending existing research that has provided a way to produce values for categorical fields. The SV approach introduced here uses machine learning techniques to derive values of numeric fields that are based on a variable number of pertinent historic messages. Our empirical evaluation demonstrates that the Cognitive SV approach can produce responses with the appropriate fields and accurately predict values of numeric fields across three data sets, some of them based on stateful protocols.

Choosing the appropriate forecasting model for predictive parameter control.

All commonly used stochastic optimisation algorithms have to be parameterised to perform effectively. Adaptive parameter control (APC) is an effective method used for this purpose. APC repeatedly adjusts parameter values during the optimisation process for optimal algorithm performance. The assignment of parameter values for a given iteration is based on previously measured performance. In recent research, time series prediction has been proposed as a method of projecting the probabilities to use for parameter value selection. In this work, we examine the suitability of a variety of prediction methods for the projection of future parameter performance based on previous data. All considered prediction methods have assumptions the time series data has to conform to for the prediction method to provide accurate projections. Looking specifically at parameters of evolutionary algorithms (EAs), we find that all standard EA parameters with the exception of population size conform largely to the assumptions made by the considered prediction methods. Evaluating the performance of these prediction methods, we find that linear regression provides the best results by a very small and statistically insignificant margin. Regardless of the prediction method, predictive parameter control outperforms state of the art parameter control methods when the performance data adheres to the assumptions made by the prediction method. When a parameter's performance data does not adhere to the assumptions made by the forecasting method, the use of prediction does not have a notable adverse impact on the algorithm's performance.