5 ms·

I would be very interested in a post mortem of the software used called SkySolver. Its supposed to be a Java Application which is said to be developed by Accent

I would be very interested in a post mortem of the software used called SkySolver. Its supposed to be a Java Application which is said to be developed by Accenture? Anyone have actual technical insights into why it failed?

Tao3300 4y ago
I'll bet it's just another excuse. The situation was probably so bad that the only solutions SkySolver could offer were beyond their capacity. There were probably partial solutions that were possible, but they rolled the dice on being able to resolve it by overworking people. They decided to double down instead of paying a whole lot of vouchers and refunds they'd rather not have to. Is it the software's fault the bet didn't pay out? As the problem snowballs and becomes more intractable, you get too far off the script and reach a point of no return. I know nothing about the SkySolver or leadership at Southwest, but that seems like a very likely scenario based on what I've seen of what corporate types expect from software and rank-and-file employees.
- ainvb 4y ago
  SkySolver is a total shit show no doubt, but it’s not the entirety of the problem. Keep in mind when you have disruptions you actually have multiple products. SkySolver only manages the crew. They have another system that manages the aircraft and flight schedule. Believe it or not these systems are mostly decoupled - the schedule is modified first, then the crew, and there may even be a feedback loop back to the schedule if there is no crew solution. Multiple things can be true here.
DLarsen 4y ago
About a decade ago, I was part of a startup trying to disrupt crew scheduling. It's a non-trivial operations problem when you aim to honor crew preferences, union-negotiated affordances, FAA legalities, etc. We were only involved in the pre-planned schedules. At that time, the airlines we were courting had entirely different human-hravybsystem to resolve real time issues. There was some level of reserve redundancy baked in so the human planners had some wiggle room to work with... but redundancy is expensive to maintain. As a relatively new engineer at the time it was a pretty neat domain with big $$$ at stake. As it turns out, pretty much nobody wanted to take the risk on a new system even if it had provably better schedules for all parties. All it takes is one snafu for the whole thing to turn into a major regret.
- kilroy123 4y ago
  Interesting. My last job was doing something similar but much more ambitious. Manage the entire fleet. All tails, flight legs, and crew. It was for a much smaller US airline but still a household name. The crew aren't union so that helped. But it was tricky to manage the crew part.
  - nlstitch 4y ago
    Im actually working at a startup that wants to use an algorithm to plan transport on large scale. ( Different Industry though). Got any tips or insights on biggest challenges?
    - DLarsen 4y ago
      Does the algo work? Is it better than the status quo and under what circumstances is it susceptible to failure? Does it know when the whole system is over-constrained? If we assume the algo rocks (because you have operations research veterans), what stage of planning does your system address? What is the experience required and cost of manual intervention? Fun problem. Human factors and edge cases abound. So much is at stake for a system that already works (to any predictable degree). I'd love to hear more though.
      - nlstitch 4y ago
        The algo works. We have two algorithmic experts onboard, of which one is a professor and one is already working with algos each day (e.g. worked on patients/beds capacity challenge when covid began). The startup focusses on assigning goods to transport. (As in its not part of the commercial sales process.. only execution of the required capacity planning that comes after it) .Something that has been tried 20 times and failed. It currently is done manually. So yeah, its going to be a challenge.
    - kilroy123 4y ago
      The biggest challenges we had were more political. Getting buy-in across the airline as well as _support_ from the airline side to integrate their data and system into ours. Sadly, it was a people problem, not a tech or algo problem.

Tao3300 4y ago

I'll bet it's just another excuse. The situation was probably so bad that the only solutions SkySolver could offer were beyond their capacity. There were probably partial solutions that were possible, but they rolled the dice on being able to resolve it by overworking people. They decided to double down instead of paying a whole lot of vouchers and refunds they'd rather not have to. Is it the software's fault the bet didn't pay out? As the problem snowballs and becomes more intractable, you get too far off the script and reach a point of no return. I know nothing about the SkySolver or leadership at Southwest, but that seems like a very likely scenario based on what I've seen of what corporate types expect from software and rank-and-file employees.

ainvb 4y ago
SkySolver is a total shit show no doubt, but it’s not the entirety of the problem. Keep in mind when you have disruptions you actually have multiple products. SkySolver only manages the crew. They have another system that manages the aircraft and flight schedule. Believe it or not these systems are mostly decoupled - the schedule is modified first, then the crew, and there may even be a feedback loop back to the schedule if there is no crew solution. Multiple things can be true here.

DLarsen 4y ago

About a decade ago, I was part of a startup trying to disrupt crew scheduling. It's a non-trivial operations problem when you aim to honor crew preferences, union-negotiated affordances, FAA legalities, etc. We were only involved in the pre-planned schedules. At that time, the airlines we were courting had entirely different human-hravybsystem to resolve real time issues. There was some level of reserve redundancy baked in so the human planners had some wiggle room to work with... but redundancy is expensive to maintain. As a relatively new engineer at the time it was a pretty neat domain with big $$$ at stake. As it turns out, pretty much nobody wanted to take the risk on a new system even if it had provably better schedules for all parties. All it takes is one snafu for the whole thing to turn into a major regret.

kilroy123 4y ago
Interesting. My last job was doing something similar but much more ambitious. Manage the entire fleet. All tails, flight legs, and crew. It was for a much smaller US airline but still a household name. The crew aren't union so that helped. But it was tricky to manage the crew part.
- nlstitch 4y ago
  Im actually working at a startup that wants to use an algorithm to plan transport on large scale. ( Different Industry though). Got any tips or insights on biggest challenges?
  - DLarsen 4y ago
    Does the algo work? Is it better than the status quo and under what circumstances is it susceptible to failure? Does it know when the whole system is over-constrained? If we assume the algo rocks (because you have operations research veterans), what stage of planning does your system address? What is the experience required and cost of manual intervention? Fun problem. Human factors and edge cases abound. So much is at stake for a system that already works (to any predictable degree). I'd love to hear more though.
    - nlstitch 4y ago
      The algo works. We have two algorithmic experts onboard, of which one is a professor and one is already working with algos each day (e.g. worked on patients/beds capacity challenge when covid began). The startup focusses on assigning goods to transport. (As in its not part of the commercial sales process.. only execution of the required capacity planning that comes after it) .Something that has been tried 20 times and failed. It currently is done manually. So yeah, its going to be a challenge.
  - kilroy123 4y ago
    The biggest challenges we had were more political. Getting buy-in across the airline as well as _support_ from the airline side to integrate their data and system into ours. Sadly, it was a people problem, not a tech or algo problem.