Civil Engineering May 2022 | Vol 30 No 4

Civil Engineering May 2022 41 time, dedicated infrastructure is shared between all qualifying public transport operators (formal and informal) in order to spread the benefits of dedicated road space more widely. In the longer run, as passenger volumes grow and the network evolves, shared lanes may graduate to full bus lanes. 3. Market behaviour: How would informal operators respond to shared infrastructure? The third study acknowledges that our ability to implement shared infrastructure is hampered by the fact that government actually has little control over the behav- iour of MBT operators, making it difficult to predict how they might use shared lanes. Questions arise as to whether MBT drivers want to migrate to shared lanes, and, if so, whether they complement or compete with the formal BRT service. To start answering such questions, we developed a tool to capture the dynamic behaviour of MBT operators using an agent-based simulation model within the Multi-Agent Transport Simulation (MATSim) framework. For a specified scenario including shared lanes along a hypothetical BRT line, MBT operators can modify their service parameters such as fleet size, routes, and frequencies in pursuit of optimal ridership and profit- ability. The passenger demand is fixed in terms of total trips, but passengers can choose between BRT and MBT services to minimise their time and cost expenditure. In this way we can predict the evolution of MBT service patterns over time in response to realistic passenger preferences and network conditions. Initial results showed that MBT operators are indeed sensitive to the decisions made by system planners. Their sensitivity is highest to the service frequency or headway of the BRT: as headways increase (i.e. fewer BRT buses per hour), MBT operators capitalise by adding more vehicles to the route and competing more strongly with the BRT (Figure 4). This might indicate the emergence of scavenging behaviour from the MBT drivers who could use their high flexibility to adjust their schedules to serve the same stops just before the arrival of BRT buses. Efforts like these to better understand interactions between informal and formal parts of the system are needed to avoid unintended consequences and devise ap- propriate incentives and regulatory tools of a hybrid system. WHERE TO FROM HERE? While these cases start to illustrate the potential for prioritising road space for bus and MBT vehicles in an integrated multimodal network, they also point to many uncertainties and knowledge gaps. Perhaps the most useful way forward will be the implementation of some small-scale pilot projects to start getting real-world results under carefully chosen conditions. These will become the basis of further development of planning approaches, design warrants, and opera- tional strategies. It is key that these efforts involve the MBT industry as knowledge partners that can help steer them in a mutually beneficial direction. ACKNOWLEDGMENTS This paper is partly based on Modelling the impact of priority infrastructure on the performance of minibus-taxi services in Southern Africa by Lourens de Beer and Christo Venter, which co-won the SAICE Transport Engineering Division’s award for the Best Paper by a Young Professional (under 35) at the Southern African Transport Conference 2021. A comprehensive paper on this topic, titled Priority infrastructure for minibus- taxis: an analytical model of potential benefits and impacts , was also published in the December 2021 (Vol.63, No.4) issue of the SAICE Journal. Other references are available on request from the authors. The research is partly sponsored by the Volvo Research and Educational Foundations through the BRT+ Centre of Excellence.  Figure 3 VISSIM model (left) of the Nana Sita BRT corridor (right), Tshwane Figure 4 Box plot showing range of MBT fleet size adjustments as BRT headway rises Headway (min) Paratransit vehicles 60 50 40 30 20 5 10 15 20