My greatest frustration is that they complain about black box algorithms when in fact it is a perfectly clear, ambiguity-free optimization function.
https://www.google.com/amp/s/www.wired.com/story/joi-ito-ai-...
Sure, when you introduce an approach that suddenly creates considerable inconvenience for some population of parents, sure they complain (no one likes having to drop off kids at ~ vaguely recalling ~ 10 am and pick at 5 pm, sheesh, it was something terrible). Some large portion of working people are fairly dependent their children having ordinary school schedules. Maybe these were relatively better off parents in the public school system - where the actually wealthy use private schools (I know Boston well enough to attest to this). So the algorithm failed to "optimize for those they could effectively screw" but hey, would it be too much to ask to just give all schools a sane schedule? Maybe pony up that extra money, give up the need to optimize everything, keep the tail of bus optimization from wagging the dog of school scheduling?
That’s a significant change in wake up time. Yeah, I’m not keen on that. It’s a non-starter.
Those early hours exist. Someone has to take them (without doubling the bus fleet size). And it's better for the students if they have the early hours in elementary and not in high school.
But fine, if this change is somehow egregious compared to other ones, give a 5 year warning.
The schedule benefits their kids, with a net zero change in the convenience of starting hours. What a shortsighted way to complain.
This is not how to build concensus for your argument.
I recently moved due to concerns outside my control, and the school time changed from 8:30am to 7:15am for my 6-year-old. This is manageable, but not a minor change at all. These types of changes are very significant for everyone in the region.
Sure, I'm not trying to convince the population being affected with that. Here's how I would say it:
"Your kids are going through 12 years of education. This change to the entire system will make it easier for them to learn, and to be smarter adults with better focusing habits. For some people the hours are tougher right now, I'm sorry, but it's not like those hours came out of nowhere. Your students were going to have school at those hours eventually. We've only changed which years."
It's not about how hard the change is. It's that a few years of early school starts are inevitable, so it's a fair trade.
Also what if your family has both teenagers and preteens? In my family we drove, so the schedule was based on when my parents had to go to work. I’d often arrive at school when the library opened or just before, which is a good hour before when classes started. Then again, my school experience wasn’t in the US, and schools started 8:30am / 8:45am consistently across the city.
"When I was a kid" preteens and teenagers were expected to be able to get to the bus stop on their own; But maybe that's just because both of my parents worked and left before the bus came.
Many of these systems start optimizing some simple criteria and work fairly well. But with additional constraints they fall apart.
Even companies that manage their own fleets for their own needs would like different things to purely optimizing distance/duration. They might want to balance the workload across all of their drivers, separate the drivers so they drive in a particular area, undefined start or end of the driving job.
In the case above, something could have definitely be done, if the algorithm was efficient enough.
Well, routing school buses is already less than trivial. You have pickup & delivery. So there's a precedence constraint to orders. Child won't be kept for hours in the bus, so the bus has to make several trips to school. Limiting the time between delivery of the child and the pickup is already less than trivial.
Just to know if route is feasible (all constraints satisfied) given a list of child pickups, child dropoffs and time limit between pickup and dropoff is nontrivial and can mess up the optimization.
Adding school shifts to the equation, minimizing number of vehicles and a bunch of other constraints might make the optimization just too slow or too constrained for an algorithm that was working incredible without all those constraints.
And 'optimization' solutions result in 1hour+ rides to a school just down the road, as the bus winds about the countryside scavenging the few remaining students. The longer the trip gets, the more that find another way, the emptier the bus, the longer the trip to try to fill it again.
We drove our kids to school for most of their school careers, even though three busses went by our house every day.
Here's [0] a paper where they analyze the mistake of the feasibility check that experts in the field failed to do properly. Here's [1] a paper aggregating all the timing problems that arise and their algorithmic complexity. Some of the timing problems, including the constraint of limiting the time of the passenger in the bus had O(n^2) or O(n^3) feasibility checks. That's slow. Especially slow if combined with integer linear programming or branch and cut algorithms.
If your system instead minimizes the riding time by adding a cost function to a constraint, making it soft, in most cases the cost function is so ill defined that the solution no longer does what you want, can hardly minimize all the constraint to a normal solution, and you get a huge mess.
There's no state of the art solution that models these constraints as soft ones.
These problems being standard does not mean that they are simple.
0: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.927...
1: https://w1.cirrelt.ca/~vidalt/papers/Timing-Problems-Final.p...
Big changes that impact tens or hundreds of thousands of people need to be staged out to minimize impact. The Big Bang approach doesn’t work for problems like this — i bet a few thousand people would lose jobs over these schedule changes.