The PeopleMover being automated, electric, and lightweight is very cheap to operate. A train is far more expensive to operate, but with a GO Train you divide those costs by far more end users, whereas an AirRail service has more staff, higher marketing costs, and additional real estate requirements and has less users to cover those costs (in addition to splitting the higher infrastructure costs amongst the users).
But you still need more trains between Union and Woodbine to make it a useful airport link. They should run every 15 or at least 20 min and this is more frequent than the present-day Brampton - Georgetown service. The cost of extra trains adds to the cost of extending PeopleMover.
No, there are GO Trains and VIA Rail there as well as the TTC. That is the whole point. It is right off the 427 so GO buses would be very practical at that location as well.
All those connections are mostly for places within GTA but outside Toronto. From within Toronto, the only two practical ways to get to Woodbine will be GO train or Finch West LRT. For their passengers, extending those lines into the airport means one less transfer when going to T1 at least (and in case of Finch LRT, perhaps T3 as well).
At what cost and how does the transfer benefit the end users which would likely out number the origins from Union station?
The cost of stopping at Woodbine is probably minimal, if the station will be added anyway.
Regarding the end users, ideally the ARL should service all stops en route. The number of transfers for groups of riders:
1) Union, plus much of Toronto south of Eglinton - to T1: ARL = 0, PeopleMover = 1
2) Union, plus much of Toronto south of Eglinton - to T3: ARL = 1, PeopleMover = 1
3) GTA via Woodbine - to T1: ARL = 1, PeopleMover = 1
4) GTA via Woodbine - to T3: ARL = 2, PeopleMover = 1
5) North-west of Toronto via Finch LRT: depends on whether Finch LRT goes beyond Woodbine.
So, there is no difference for groups 2 and 3. Group 1 is better off with ARL, group 4 with PeopleMover; but 1 likely outnumbers 4.