Bus planning from a modelling approach

Figure 1: Functions of the Bus Planning Model

Figure 1: Functions of the Bus Planning Model

Figure 3: Planning process of network evaluation and optimisation

Figure 3: Planning process of network evaluation and optimisation

Figure 4: A simple network for illustration

Figure 4: A simple network for illustration

Figure 2: The BPM feeds outputs to the AVM and AFC systems to create a loop

Figure 2: The BPM feeds outputs to the AVM and AFC systems to create a loop

The Bus Planning Model (BPM) from the Department of Transport (DoT), Emirate of Abu Dhabi, United Arab Emirates, is part of its integrated bus planning and operating systems. It uses actual bus operations data captured by Automated Vehicle Monitoring (AVM) and Automated Fare Collection (AFC) Systems to facilitate bus planners in the evaluation and optimisation of bus networks and bus operational planning. By Oscar Jiang Jing

The DoT has developed its bus planning model to provide reliable and efficient bus networks to cater for increasing travel demand. A cost model was developed as a module of the Bus Planning Model (BPM) for network optimisation, budgeting and cost recovery plans. The model was built with input from the DoT’s strategic transport demand model, street network, bus network supply data and bus operations data. It provides a means for network evaluation and develops plans of network changes and improvement. The plans include detailed and specified bus routes, timetables and number of vehicle blocks. The model is also able to develop long-range plans to evaluate the needs of future bus fleet and subsidy requirements.

The BPM takes data from the AVM and AFC systems of bus travel time between bus stops and between termini, passengers’ boarding and alighting patterns, passenger load and travel origins-destinations. Based on the data, the BPM carries out network performance review, network design and timetable planning. The BPM feeds the final planning output to the scheduling system for refined vehicle blocks and duty planning, which are to be passed on to the AVM and AFC systems to create a loop (Figure 2).

One of the key functions of the BPM is network evaluation and optimisation. The network is evaluated based on five Key Planning Indicators (KPIs), namely coverage, crowdedness, utilisation, speed and transfer. Alongside network evaluation, potential demand in areas where there are no bus services is also considered. Potential demand can be estimated either from the strategic transport demand model or taxi data. Based on network evaluation and potential demand, bus planners propose network changes, within the constraints of bus availability. The model evaluates the planned network, enabling planners to test different scenarios of network changes until the optimum choice is apparent. 

Operations planning

Bus timetables need to be carefully planned. Passengers expect timely and regular services, while the DoT aims to minimise unnecessary delays and limit excess capacity. Bus operators need to adherto timetable targets in variable and dynamic traffic conditions. Runtime profiling is the key procedure in timetabling. Runtime is defined as the travel time of a bus from one terminal to another, and is the deciding factor behind the number of vehicles required for the network. It has a direct effect on service reliability and timetable adherence. Not only is it of key importance, it is also highly complex. Runtime varies across the day in peak and off peak periods, and on different ‘day types’.

A specific procedure – 24-hour profiling (for each day type) – was developed to solve the issues inherent given the complexity discussed above, and to set runtime for each time profile of each service without conflict (see Figure 4). The network consists of 3 lines, 9 bus stops (including termini) and 9 links. The travel time of all buses of each link for at least one month’s period is recorded by the AVM system over 24 hours and averaged for each hour (one hour is one runtime profile). Subsequently, a 24-hour link time matrix, which contains all links of the network, is created. 

The runtime of each runtime profile of each bus service can be summed up with the travel time (including stop time) of each link along the route course of the corresponding hour of the runtime profile. 

RuntimeHi (Line 1, West-East) = L1Hi + L4Hi +L5Hi + L6Hi

RuntimeHi (Line 2, WE) = L2Hi + L4Hi + L8Hi + L9Hi

RuntimeHi (Line 3, WE) = L3Hi + L7Hi + L9Hi

Where, 

Hi – Profile of Hour i;

i – from 0 to 23.

Furthermore, the accumulated runtimes are compared with the entire journey times recorded by the AVM system. If the difference is more than 5 minutes, the accumulated runtimes are adjusted. These processes are automated in the BPM and applied to the actual bus network. With runtimes and journeys created, which are associated with departure times, the BPM generates vehicle blocks.

Cost model

Operating cost is one of the key outputs from the BPM after vehicle journeys and vehicle blocks are created. The operating cost consists of four components: fuel consumption, operators’ fees, insurance and maintenance. Cost rates are set in the model. The operating cost is calculated by the model through pre-set calculation procedures, which can be exported to Excel. The output can be used for network optimisation, budgeting and cost recovery plans.

Oscar Jiang Jing, Lead Planner, Transport Planning, Department of Transport, Abu Dhabi