The Pitstop

@BobCross it's the nature of the problem: try to find the optimal ordering of a list of 60 items. Certain orderings will "perform" better than others. It helps to know all possible orderings in advance so that an algorithm can pick and choose what it sees fit

Bob Cross

@Zaid Optimal? Or in the top 10%? Those are two different problems but the second one is usually much easier to solve

Zaid

6:02 PM

@BobCross the goal is to get the best performing permutation but naturally, exploring all 60! orderings is impossible

The fun part is leveraging domain knowledge to reduce the complexity of the problem. Heuristics for the win!

Bob Cross

With an A Star algorithm in there somewhere?

Zaid

Interesting, never heard of that algo

Bob Cross

A* search algorithm

In computer science, A* (pronounced as "A star" ( listen)) is a computer algorithm that is widely used in pathfinding and graph traversal, the process of plotting an efficiently traversable path between multiple points, called nodes. Noted for its performance and accuracy, it enjoys widespread use. However, in practical travel-routing systems, it is generally outperformed by algorithms which can pre-process the graph to attain better performance, although other work has found A* to be superior to other approaches. Peter Hart, Nils Nilsson and Bertram Raphael of Stanford Research Institute (now...

Zaid

Oh, it's an extension of Dijkstra

Bob Cross

It's really old

Zaid

6:07 PM

I don't see this as a traversal problem

But hmmm, it's an interesting way to think of it

akadian

@BobCross Aren't you an Oil guru?

Bob Cross

@akadian don't confuse me with Bob the Oil guy - that's a really different thing

2

@Zaid you can use it to score permutations

akadian

I know, but weren't you bragging about your Oil badge or something before?

Bob Cross

if you're traversing the permutation space, you can say that "if my score is already this bad, this isn't a valid combination"

@akadian ... yes, those are true things.... ;-)

akadian

@BobCross Any chance you can help?

3

Q: Effects of higher oil viscosity

After reading "Can viscosity affect gas mileage?", it appears that it can affect gas mileage. However, what would be the consequences of going from the manufacturer's recommended 0w-20 to the 5w-40 that fansites often say is "better for tracking" on my car? Will it affect the wear on my engine? ...

I like the current answer, but I'm waiting for a reply on my comment before I accept. Unless more answers come up.

Zaid

6:14 PM

@BobCross all sequences are valid, but some are more desirable than others

DucatiKiller

@akadian He wasn't bragging. He was simply ensuring everyone knew how knowledgeable he was about oil.

Bob Cross

@Zaid right, so you're assigning a fitness score

akadian

@DucatiKiller Can't believe I confused the two!

Zaid

@akadian you should check with @Pᴀᴜʟsᴛᴇʀ2

DucatiKiller

blackflag.jalopnik.com/…

akadian

6:16 PM

Oh and I'm going to Vermont/New York in 2 weeks. Any recommendations on a nice road to drive on?

Last 2-3 weeks before I swap to winter tires... Got to do a road trip before! :D

Zaid

@BobCross let me think about this a bit. I can see potential with something like A* or Dijkstra but the concern is that it would get "stuck" with a certain main path and not consider other "wildly different" paths that could be more competitive.

Bob Cross

That's where it helps to redefine the problem as "very good" and not "best"

It's sort of like redefining everything as a path-finding problem

Eventually, you need a path

Zaid

I get that. Given that the baseline is given by hand-calculated trial and error, I'd say the higher-ups will be over the moon with "very good" :)

Bob Cross

@Zaid See? I helped! With a thing! :-)

Zaid

@BobCross would you happen to know what this problem is called? I want to group these 60 items in pairs, such that a property of the pair across all pairs is as equal as possible. It's not k-means clustering, I think

@BobCross I'm a big fan of elegant solutions :)

Bob Cross

6:30 PM

But the pairs are disjoint?

Zaid

Each pair is independent of other pairs, if that's what you're asking

Bob Cross

I know what you mean but I can't think of the terms right now.

It's not far away from a nearest neighbor problem:

Nearest-neighbor chain algorithm

In the theory of cluster analysis, the nearest-neighbor chain algorithm is a method that can be used to perform several types of agglomerative hierarchical clustering, in which a hierarchy of clusters is created by repeatedly merging pairs of smaller clusters to form larger clusters. In particular it can be used for Ward's method, complete-linkage clustering, and single-linkage clustering, which all work by merging the closest two clusters under different definitions of the distance between clusters. The main idea of the algorithm is to find pairs of clusters to merge by following paths in the...

Sounds like you kind of don't need the chain issue but might get that for free

Or maybe you do need the chain

In computability theory, we used to do a lot of transforming problems into other problems so that we could estimate run time

Zaid

@BobCross funny you should say that. We identified that as one of the uses of this pairing idea is to establish a theoretical worst-case scenario

Bob Cross

Okay, sounds like we're in roughly the same conceptual space. Hopefully, this shook loose some ideas. Sometimes it helps to see if you can figure out a lower bound for the shortest path (e.g., classic New York to LA problem - path can't be shorter than great circle segment between them).

If you can come within X% of that lower bound, you know that you have something quite good and can stop searching.

Depends on your application. Nerd nerd nerd.

8:14 PM

Actually, I'm not sure I like the design of the bolt? There'd be a large stress riser at the center just above the threads. Anyway ...

DucatiKiller

@Pᴀᴜʟsᴛᴇʀ2 Yeah, those guys probably got it all wrong. lol

Pᴀᴜʟsᴛᴇʀ2

@DucatiKiller - Oh, absolutely ;-)

Bob Cross

Okay, that's a cool video

DucatiKiller

8:29 PM

@Pᴀᴜʟsᴛᴇʀ2 They should be coming here and consulting with us before they just go and start designing bolts.

Pᴀᴜʟsᴛᴇʀ2

9:12 PM

@DucatiKiller - Um ... now that you say it, yes they should!

Pᴀᴜʟsᴛᴇʀ2

9:26 PM

This is not what I thought this meant ...

Myself

@Pᴀᴜʟsᴛᴇʀ2 Great :)

How are things, paul?

Pᴀᴜʟsᴛᴇʀ2

It's okay. I've been thinking too much and it's starting to hurt.

I need to go out and work on the Civic.

I got it started this afternoon. I didn't see any leaks out of the hydraulic lines, so I think I can put it back together now.

Myself

I need to fix a corsa

O2 sensor, exhaust and all the shocks

I thought about doing it by the book and to replace the strut-bolts

When I ordered them at the dealership they looked at me like crazy

Delivery time for the bolts: 4 working days.

I guess I am the only one who's replacing the bolts there

Pᴀᴜʟsᴛᴇʀ2

Well, in most cases you cannot go wrong by the book ... in most cases.

New bolts won't hurt anything even if you didn't need to replace them.

Myself

9:44 PM

Is something still to do on the civic? Except putting the rest together?

Pᴀᴜʟsᴛᴇʀ2

I think just putting it back together. I'll probably need an alignment ... I'll get the cheapest one I can throw at it, though.

Might just see how it drives and leave it be. It'd only be the toe, so if it tracks well and doesn't pull, I might just leave it alone.

I was pretty sure to measure everything and it looks pretty close via a tape measure.

Myself

"Tape measure"?

Pᴀᴜʟsᴛᴇʀ2

Myself

lol, I was thinking about some advanced technique

Pᴀᴜʟsᴛᴇʀ2

10:06 PM

Nah ... about as advanced as it gets around my house ;-)

I'm going to work on the civic tomorrow ... I'm just not feeling it tonight.

Myself

gotta go, cya

SteveRacer

11:41 PM

@DucatiKiller Not sure about the groove above the threads. It's a "Camber Rod" bolt, and mounted in double shear, so maybe the groove is for lubrication. It's also unsprung weight, which might explain why the threads are cut into some sort of alloy, rather than rolled. I don't think this is a high-tensile fastener on par with AN or aerospace applications. I saw no heat treatment. I don't think it's there for clamping, just to act as a pivot - hence the highly accurate shank diameter.

I couldn't read the material spec on the print, but it looked to cut like a high grade 10xx alloy. Titanium would make more sense as unspung weight savings. Whatever the material composition, I am certain it contains equal parts unobtanium and expensonium.

Transcript for