4 ms·

Problem is that academics are rarely experts at programming or have knowledge of computer architectures as much as someone in the industry. There are various tr

Problem is that academics are rarely experts at programming or have knowledge of computer architectures as much as someone in the industry. There are various tricks that are never taught at college, therefore academics have no idea some stuff even exists. Best example is discrete optimization research (traveling salesman, vehicle routing and its variants, schedule rostering etc.). Stuff you find in the papers there achieves state-of-the-art results very slowly (using integer linear programming or some rarely optimized heuristics) making you believe these instances of a general NP-hard problem can't be solved quickly. When you start tinkering, you either find that data structures can be added that reduce the complexity significantly or that there are regularities in instances that, when exploited, support massive speedups. I would say that TSP research is an exception but most of stuff coming out that has a lot citations is way too slow and is never as brilliantly implemented as Lin Kernighan heuristic or other stuff from the age of insanely slow computers.

raverbashing 5y ago
> Stuff you find in the papers there achieves state-of-the-art results very slowly (using integer linear programming or some rarely optimized heuristics) making you believe these instances of a general NP-hard problem can't be solved quickly. Yes, or looking for things like mathematical purity, linear problem statement, etc In practice: you don't need the best solution and you can get a great solution in a multitude of ways.
pyrale 5y ago
From what I saw (I worked in an academia-consulting partnership to solve scheduling problems in transports), the OR researchers very frequently work in association with OR consulting shops, and they're well-aware of the difference between academic datasets and industrial datasets. In the conferences I saw, it was not infrequent to see different tracks for industrial and academic papers, both attended by the same crowd. The point I agree with, though, is that this is not reflected in the papers. Academic papers focus on academic instances because they are more general (and usually harder, as you said) and because optimizations of specific instances of the problem are not that useful from an academic pov. It's hard to know who works with who and who has experience with what if you're not an insider, though.
makmanalp 5y ago
> Problem is that academics are rarely experts at programming or have knowledge of computer architectures as much as someone in the industry. There are various tricks that are never taught at college, therefore academics have no idea some stuff even exists. I want to push back on this generalization a bit. The academics that are focused on pushing the mathematical boundaries of discrete optimization are focused, no surprise, on only that. If theoretical improvements is not what you want, don't read those papers, read ones what people more focused on applications. Read literally any databases paper, or stuff like hyperscan, simdjson. I'd argue that a non-trivial amount of these are vastly /ahead/ of what's standard in industry, but industry is slow to adapt new ideas and catch up because of legacy software and existing codebases. Very similar stuff in the programming languages sphere, it took ages for industry to adapt ideas that were popular in academia for a long time (eg: Rust, LINQ). The idea that academia (at least in CS) is an ivory tower, far from the concerns of industry, is not very true as of recent. There's a lot of cross pollination and a revolving door of people going back and forth from one to the other spreading ideas.
- knuthsat 5y ago
  I must say that when it comes to discrete optimization, the genetic/ant/simulated annealing/etc. stuff is more popular in academia than in industry (at least the industry that doesn't heavily include academics). Works like Lin-Kernighan heuristic are extremely rare and a bunch of knowledge exists in industry only. Even the mentioned heuristic was for decades being implemented incorrectly until one individual came and demonstrated its superiority (K. Helsgaun). I mean, most of the code that gets released today, doesn't even handle time windows constraint in the best way possible (optimal updates, constant time feasibility evaluations etc.). I believe all open source vehicle routing libraries do not have any data-structure relevant optimizations for handling time windows, task precedence, dynamic workshift starts etc. All are fairly simple implementation tricks that just got lost under giant amounts of population based heuristics or mathematical linear programming approaches.
  - jhgb 5y ago
    > Even the mentioned heuristic was for decades being implemented incorrectly until one individual came and demonstrated its superiority (K. Helsgaun). Does this mean that the Linkern program in the Concorde TSP suite is also implemented incorrectly?
kaba0 5y ago
You do realize there is a whole area of the research field dedicated for heuristic algorithms? They have a proper academic basis just as much as the “correct” solutions.
- jhgb 5y ago
  Since he mentions the L-K heuristic, he obviously does.

raverbashing 5y ago

> Stuff you find in the papers there achieves state-of-the-art results very slowly (using integer linear programming or some rarely optimized heuristics) making you believe these instances of a general NP-hard problem can't be solved quickly. Yes, or looking for things like mathematical purity, linear problem statement, etc In practice: you don't need the best solution and you can get a great solution in a multitude of ways.

pyrale 5y ago

From what I saw (I worked in an academia-consulting partnership to solve scheduling problems in transports), the OR researchers very frequently work in association with OR consulting shops, and they're well-aware of the difference between academic datasets and industrial datasets. In the conferences I saw, it was not infrequent to see different tracks for industrial and academic papers, both attended by the same crowd. The point I agree with, though, is that this is not reflected in the papers. Academic papers focus on academic instances because they are more general (and usually harder, as you said) and because optimizations of specific instances of the problem are not that useful from an academic pov. It's hard to know who works with who and who has experience with what if you're not an insider, though.

makmanalp 5y ago

> Problem is that academics are rarely experts at programming or have knowledge of computer architectures as much as someone in the industry. There are various tricks that are never taught at college, therefore academics have no idea some stuff even exists. I want to push back on this generalization a bit. The academics that are focused on pushing the mathematical boundaries of discrete optimization are focused, no surprise, on only that. If theoretical improvements is not what you want, don't read those papers, read ones what people more focused on applications. Read literally any databases paper, or stuff like hyperscan, simdjson. I'd argue that a non-trivial amount of these are vastly /ahead/ of what's standard in industry, but industry is slow to adapt new ideas and catch up because of legacy software and existing codebases. Very similar stuff in the programming languages sphere, it took ages for industry to adapt ideas that were popular in academia for a long time (eg: Rust, LINQ). The idea that academia (at least in CS) is an ivory tower, far from the concerns of industry, is not very true as of recent. There's a lot of cross pollination and a revolving door of people going back and forth from one to the other spreading ideas.

knuthsat 5y ago
I must say that when it comes to discrete optimization, the genetic/ant/simulated annealing/etc. stuff is more popular in academia than in industry (at least the industry that doesn't heavily include academics). Works like Lin-Kernighan heuristic are extremely rare and a bunch of knowledge exists in industry only. Even the mentioned heuristic was for decades being implemented incorrectly until one individual came and demonstrated its superiority (K. Helsgaun). I mean, most of the code that gets released today, doesn't even handle time windows constraint in the best way possible (optimal updates, constant time feasibility evaluations etc.). I believe all open source vehicle routing libraries do not have any data-structure relevant optimizations for handling time windows, task precedence, dynamic workshift starts etc. All are fairly simple implementation tricks that just got lost under giant amounts of population based heuristics or mathematical linear programming approaches.
- jhgb 5y ago
  > Even the mentioned heuristic was for decades being implemented incorrectly until one individual came and demonstrated its superiority (K. Helsgaun). Does this mean that the Linkern program in the Concorde TSP suite is also implemented incorrectly?

kaba0 5y ago

You do realize there is a whole area of the research field dedicated for heuristic algorithms? They have a proper academic basis just as much as the “correct” solutions.

jhgb 5y ago
Since he mentions the L-K heuristic, he obviously does.