Then just increase the number in the utility function that I mentioned represents the utility of living in the same place for periods longer than a month. Are you asking about periods longer than that? I guess there’s no reason the system couldn’t look even further back. A move penalty that changes the longer they’ve been in that position.
I’ve considered that, yeah, public incremental improvement processes wouldn’t work for everyone. Likely it would have to go away after the initial stages. I’d originally planned for a process where the preference expression data would never leave city hardware, where solvers would instead send their optimization programs into city hardware, which would pass the data to the program, run for an hour or whatever, pass out the resultant solution, then reset.
Yeah, I don’t feel like that really covers it. Maybe what I meant wasn’t really utility per se, but rather an intuition about people ragequitting this system not just because of the moving issue but because of a lot of little unpredictable annoyances adding up, with moving into a new space being one example of how this could happen. It seems like the more things change, the more unexpected annoyances are likely to pop up, whether within a living space, a neighborhood, the whole city, or whatever.
Like, a lot of people move one month (it could even be moves that optimize the expected utility of genuinely everyone in terms of things like proximity to friends and what kind of living space they’re assigned and so on) but suddenly after the move there are traffic jams or public transport is suddenly overwhelmed at certain times or from certain locations because when change happens at a normal rate people adapt and maybe change their route when it starts to get a little crowded or change what time they go to work or whatever, but when there’s a sudden change, there’s no time to adapt, you just have to deal with yourself and everyone else having habits and practices that may suddenly no longer make sense.
(This is just an example. I feel like there are all kinds of things that can go wrong when people change their habits suddenly. See also: covid + toilet paper.)
Maybe some sort of generally applicable habits and practices could develop, but maybe not. It seems unlikely that generally applicable habits would be as efficient as habits and practices that have had time to get optimized through use.
(Also, lots of people moving at the same time once a month is not a great way to utilize moving companies. But if they don’t move at the same time...how far from optimal can it get because of moving delays?)
I feel like this idea needs a Bill of Rights to assure residents that there is some maximum to the ways that they can be badly treated for the sake of others (only required to move once a year against their will, for example), and maybe some way to make sure that change happens somewhat incrementally, at a rate that doesn’t overcome the ability of the residents to adapt.
Maybe some sort of generally applicable habits and practices could develop, but maybe not. It seems unlikely that generally applicable habits would be as efficient as habits and practices that have had time to get optimized through use.
The word “Antifragile” springs annoyingly to mind. Constant weak shocks with lots of survivors. Maybe congestion prediction systems will tend to emerge? (Maybe those responses could be incorporated back into the utility function!?)
Also, lots of people moving at the same time once a month is not a great way to utilize moving companies. But if they don’t move at the same time...how far from optimal can it get because of moving delays?
Good thoughts. I think I can see a resolution. Moves could planned to take place throughout the month after the solution is proposed.
It occurs to me now that the optimizers are going to be a bit more complex than I’d imagined. They can’t just produce a mapping from residents to locations. Moves have to be ordered. The chains have to start at an unoccupied location and end before the month is out.
Move plans could take the capacity of the moving contracters into account, if those are known quantities.
Bill of Rights
Yeah. While this project is making me realize that going without firm guarantees is sometimes really useful (saying, “I don’t know how good it will be” enables it to become unexpectedly good. Giving it broad latitude lets it compromise on things that turn out to be more costly than was anticipated.), and I think the optimizer might end up being pretty reliable, I think there will need to be quite a few firm guarantees.
Then just increase the number in the utility function that I mentioned represents the utility of living in the same place for periods longer than a month.
Are you asking about periods longer than that? I guess there’s no reason the system couldn’t look even further back. A move penalty that changes the longer they’ve been in that position.
I’ve considered that, yeah, public incremental improvement processes wouldn’t work for everyone. Likely it would have to go away after the initial stages.
I’d originally planned for a process where the preference expression data would never leave city hardware, where solvers would instead send their optimization programs into city hardware, which would pass the data to the program, run for an hour or whatever, pass out the resultant solution, then reset.
Yeah, I don’t feel like that really covers it. Maybe what I meant wasn’t really utility per se, but rather an intuition about people ragequitting this system not just because of the moving issue but because of a lot of little unpredictable annoyances adding up, with moving into a new space being one example of how this could happen. It seems like the more things change, the more unexpected annoyances are likely to pop up, whether within a living space, a neighborhood, the whole city, or whatever.
Like, a lot of people move one month (it could even be moves that optimize the expected utility of genuinely everyone in terms of things like proximity to friends and what kind of living space they’re assigned and so on) but suddenly after the move there are traffic jams or public transport is suddenly overwhelmed at certain times or from certain locations because when change happens at a normal rate people adapt and maybe change their route when it starts to get a little crowded or change what time they go to work or whatever, but when there’s a sudden change, there’s no time to adapt, you just have to deal with yourself and everyone else having habits and practices that may suddenly no longer make sense.
(This is just an example. I feel like there are all kinds of things that can go wrong when people change their habits suddenly. See also: covid + toilet paper.)
Maybe some sort of generally applicable habits and practices could develop, but maybe not. It seems unlikely that generally applicable habits would be as efficient as habits and practices that have had time to get optimized through use.
(Also, lots of people moving at the same time once a month is not a great way to utilize moving companies. But if they don’t move at the same time...how far from optimal can it get because of moving delays?)
I feel like this idea needs a Bill of Rights to assure residents that there is some maximum to the ways that they can be badly treated for the sake of others (only required to move once a year against their will, for example), and maybe some way to make sure that change happens somewhat incrementally, at a rate that doesn’t overcome the ability of the residents to adapt.
The word “Antifragile” springs annoyingly to mind. Constant weak shocks with lots of survivors. Maybe congestion prediction systems will tend to emerge? (Maybe those responses could be incorporated back into the utility function!?)
Good thoughts. I think I can see a resolution. Moves could planned to take place throughout the month after the solution is proposed.
It occurs to me now that the optimizers are going to be a bit more complex than I’d imagined. They can’t just produce a mapping from residents to locations. Moves have to be ordered. The chains have to start at an unoccupied location and end before the month is out.
Move plans could take the capacity of the moving contracters into account, if those are known quantities.
Yeah. While this project is making me realize that going without firm guarantees is sometimes really useful (saying, “I don’t know how good it will be” enables it to become unexpectedly good. Giving it broad latitude lets it compromise on things that turn out to be more costly than was anticipated.), and I think the optimizer might end up being pretty reliable, I think there will need to be quite a few firm guarantees.