Many years back I had to read a book for uni by former SAS Airline CEO, Jan Calzon, I believe its "moments of truth". He lead SAS through an amazing turnaround in the 1980's.
One of his strategies was, never sell a seat below cost, because who are you selling it to? The annual holiday maker who's business is not adding to your bottom line and unlikely to yield repeat patronage, such last minute bookings are usually driven by people doing as you say, grabbing a short trip on a moments notice because its cheap and will not be loyal to any one brand, although V/line has no direct modal competition.
V/line seats are already below cost, so discounting further is probably not going to improve revenue to the govt. I don't know, maybe it will because a planes costs are very different with fuel at 20% and weight a big factor in fuel burn. A pax train doesn't even carry 1/3 its own weight in people.
Maybe's Jan's theory doesn't apply here. I suppose you need to look at the cost benefit analysis of doing so.
When they say "below cost", they mean below marginal cost - that is "if this extra passenger travels, how much extra will it cost us".
If you have spare seats, then marginal cost is nothing like the average cost for both planes and trains (but more so for trains).
(If you have no spare seats, then your marginal cost becomes very high, because you have to run an extra service to provide that single extra seat.)
If you have already committed to providing the service, you have spare seats, and there is no chance of you selling the seats to someone that will pay more, then it is worth selling the seat to the cheap-skate holiday maker even if its revenue is just above the marginal cost.
When you say "V/line seats are already below cost" you are referring to average cost. I would still expect the marginal cost of a passenger to be less than their fares even for V/line.
A business is not sustainable if, on average, it is selling seats for fares that are less than the average cost of each seat, where the average cost includes operating cash costs and some allowance for capital costs.
A typical approach to yield management is to estimate well ahead of running the service what sort of rmarket there might be for the service, and then dividing up the capacity of the service into fare types based on that estimate. People who don't care for flexibility or other frills that might be offered with the more expensive fares will naturally purchase the cheaper types of fare if they are available. It is because the number of seats available in those cheaper fare buckets is limited that fares then tend to rise as the time of operating the service approaches. Some people, even a relatively long time in advance, want the flexibility or other value adds associated with the premium fares (often because they are not paying for the ticket), so they still buy the expensive fare.
If the operator thinks that a service will be full even if all seats are sold near their premium fare level, then they make very few discount fares available for that service, even if you are still a long time out from the service operating. As you get closer to the time that the service is planned to operate, if the sales are not running as expected, then you might transfer additional seats from the premium fares to discount fares (or vice versa) to try and stimulate demand (or take advantage of unforeseen demand).