Here is my proposal to restructure the bus services in Singapore.
We now have 375 bus services in Singapore using 4,621 bus stops located on average 0.4 km apart. A typical bus service has 50 stops. The travelling time between each stop is about 2 minutes. A typical bus journey from terminal to terminal may take about 100 minutes or nearly 2 hours. (Some of the figures are based on my guess).
Under the proposed system, we will convert 900 bus stops to be express stops served by the express services. These express stops are located every 2 km apart. The express services stop at the express stops only. This will reduce the travelling time and can make better use of the dedicated bus lanes.
We will probably need only 80 express services to provide connections to the 900 express stops. Although the remaining 295 service can be feeder services, it is possible to combine them and reduce the number. I guess that 160 feeder services should be sufficient to cater to the needs of commuters.
With a 40% reduction in the number of services, the waiting time to take an express or feeder bus is likely to be reduced by at least 40% as there will be more buses for each service. The waiting time for an express bus can be reduced from 10 minutes to 6 minutes and for a feeder bus from 5 to 3 minutes.
The overall travelling time can probably be reduced by 30% under the new system as the express service will make less stops and the intermediate stops will be served by feeder services. Although this may require 1 or 2 more transfers, the waiting time may not be increased as the buses will arrive more frequently.
Planning the new system
We have the data of the travelling pattern to plan the new system. The data shows the number of people boarding and alighting at each bus stop during each time segment (say, every half hour). If we group the bus stops according to the nearest express stop, we will be able to measure the estimate for each of the express services for different times of the day. The travel time of the buses can be planned accordingly.
The planning can be made dynamic with the travel data for each week used to plan the scheduling for the following week.