[ home / rules / faq ] [ overboard / sfw / alt ] [ leftypol / edu / labor / siberia / lgbt / latam / hobby / tech / games / anime / music / draw / AKM ] [ meta ] [ wiki / shop / tv / tiktok / twitter / patreon ] [ GET / ref / marx / booru ]

/gulag_archive/ - Gulag Archive

Name
Options
Subject
Comment
Flag
File
Embed
Password(For file deletion.)

Not reporting is bourgeois


 

Linear programming has proven to be a feasible way of maximizing outputs given certain inputs, but it can't, for example, decide whether people would be better served building more cargo ships, more tractors, or more consumer cars using the same metal. It cannot decide how to organize the use of said materials(like for example, build one big plant or several smaller ones). How does socialism manage opportunity cost? What mechanisms under socialism(cyber or otherwise) are in place to supplement this shortcoming?

If I had to guess, I would say that this mechanism would be filled in for by the Mass Line as Maoists swear by if not a pseudo-market like Cockshott proposes. I would appreciate recommendations of further material on this topic(though I will say that the people in /cybersoc/ were extremely helpful with their takes on the topic, and I cannot deny their critique of Hayek, which compels me to study further into leftism)(Spam IP)(merge into chatgpt general)

>>2022
>but it can't, for example, decide whether people would be better served building more cargo ships, more tractors, or more consumer cars using the same metal.
Yes it can.
>It cannot decide how to organize the use of said materials(like for example, build one big plant or several smaller ones).
Yes it does.
>How does socialism manage opportunity cost?
Opportunity cost is a capitalist fiction capitalist enterprises don't directly deal with either.

Linear programming can be tasked to maximize a vector of production using a given set of available resources and methods of production. This does include multiple construction options and multiple usages of the same resources.
Perhaps you ought to actually understand linear programming before talking actual bullshit.

Here, go read this, understand it, then come back to us.

File: 1691257023170.png (100.77 KB, 736x416, ClipboardImage.png)

Like clockwork.

>>2022
Do you even understand the math behind that? Can you even understand the basics of algebra and linear algebra?

>>2025
if he's your normal austro-libertarian who only reads mises because he cant understand enough math to learn standard neoclassical econ… probably not

>>2023
OP has already been recommended this paper in the cybersoc thread

>If I had to guess, I would say that this mechanism would be filled in for by the Mass Line as Maoists swear by if not a pseudo-market like Cockshott proposes.
You could probably squish a Stakhanovite dogfooding product development cycle in there.

Dogfooding ie. eating your own dogfood if that's what you produce.

>it can't, for example, decide whether people would be better served building more cargo ships, more tractors, or more consumer cars using the same metal
That's where you're wrong kiddo
>How does socialism manage opportunity cost?
By managing it. Also by staying away from all constraints.

>>2027
Perhaps he should read it.

>>2022
> but it can't, for example, decide whether people would be better served building more cargo ships, more tractors, or more consumer cars using the same metal.
It can. Thats litterally all linear programming is used for in corporations btw. Calculate how best to use your labour and machines and resources to maximize profit.

>>2027
I apologize for not reading it earlier before making this thread. Assuming I have not exhausted your patience, I would like you to explain the workings of this graph to me. Are tops and bottoms the hypothetical stand-in for product A and product B, or do they mean something else entirely that I missed? The plan ray is just an x = y line as far as I understood, and the individual lines being segments of 3y = 2x, y = 2x, and 8y = 3x respectively.
>>2025
I will have to admit that I have not studied linear algebra as of yet; I've heard that Axler's book covers the topic pretty well. Would you recommend something different?
>>2024
That is how I've acted so far by failing to read what I was given before making a full thread.

>>2032
I don't think I do understand the graph after all. You're supposed to draw a polygon out of all the slope lines, but how does that indicate…oh, is it like you use the other points of production as shown directly above and below where the plan ray intersects? I think that's a bad explanation, but help me.

>>2032
>>2033
The graph is a representation of a function.
Hope this helped.

File: 1691337153097.png (92.48 KB, 583x251, ClipboardImage.png)

>>2032
Tops and bottoms are two products yes.
Kantonovichs diagram only works with two end products (linear programming algorithms can work with any number, but for a graph on paper in 2d its just 2).
You start with the one best at producing bottoms, then the next best, putting their line of possible production at max intensity end to end. This way it stays convex. Then you draw a line with an angle proportional to the proportion of the output you want. All things left of the line should make bottoms all things right of the line should make tops. The line intersecting allows you to calculate the amount of the sliced one to split between the two.

But all in all, it simply shows you, regardless of if you understand how linear programming works (god knows I dont understand the cutting edge algorithms), that it can determine, using a single input, and multiple possible production methods, how to maximize an output.

>>2032
Here’s the book. It has proofs

>>2022
>but it can't, for example, decide whether people would be better served building more cargo ships, more tractors, or more consumer cars using the same metal.

It can.

Problem is not that but property ownership and sabotage when planning does not benefit wannabe porkies

>>2033
Assume that you need 1 top and 1 bottom widget to get a finished product.
You have three options for producing them: auto turret lathes, turret lathes, and milling machines.
The slopes of the three machines indicate how efficiently they produce each type of thing.
The plan ray intersects with the arc to show you how much time should be spent on each type of machine.
The intersection here shows that the most efficient way to produce widgets is to set 100% of the auto turret lathe time to producing tops, 100% of the turret lathe time to the bottom, and about 90% of the milling machine time to producing bottoms.

I hope that helps.

Actually, I scanned and edited Cockshott's entire latest book (that includes that graph).
I would recommend having some experience with graphs and math for resolving problems in either science or engineering field, but it's pretty basic if you have that.

>>2039
oh fark anon, can you repost this in the cybersoc channel?

>it can
missing the point. let's look at what op said
>but it can't, for example, decide whether people would be better served building more cargo ships, more tractors, or more consumer cars using the same metal.
A sufficiently advanced algorithm can decide this. But his is all predicated on what exactly you mean by better served. Who gets to decide what the calculator thinks better served means? The ECP isn't about actually calculation at all, it's about information. /cybersoc/ actually tries to solve this but the standard response of something something linear algebra doesn't't cut it.

>>2022
Love how ancaps think the market and capitalism isn't also doing calculations just like this, except in a paradigm of less I formation, less standardization of what optimality means, and more perverse incentives to plan inefficiently by ignoring externalities and straight up minimizing benefit it gives to workers

>>2040
you mean the thread? Yeah

>>2041
>Who gets to decide what the calculator thinks better served means?
What the fuck are you talking about.
It maximizes or minimizes some output given a list of constraints. There is no ambiguity or matter of opinion. It literally calculates the best possible configuration given the information it has.
The vector of products you decide upon comes from political and economic considerations from other systems, but this is unrelated to linear programming and not a weakness of it. The economic calculation problem has nothing to do with what op is talking about.

>>2044
>It maximizes or minimizes some output
Who decides what outputs to prioritize?
>given a list of constraints.
Who defines the constraints?
> It literally calculates the best possible configuration
Define best possible configuration.
The capitalists have a simple answer on how to answer all of this, market signals.
>The economic calculation problem has nothing to do with what op is talking about.
It does because people like to pretend that linear algebra paper refutes ECP when it really doesn't.

>>2045
>Who decides what outputs to prioritize?
The planners dummy. Buy this has nothing to do with the algorithm or the "ecp"
>It does because people like to pretend that linear algebra paper refutes ECP when it really doesn't.
The only thing the ecp actually has any claim is exactly the time complexity of the calculation of a whole economy. Getting the information is no issue, it's available in the companies so getting it or better is a no brainer.

>>2045
>It does because people like to pretend that linear algebra paper refutes ECP when it really doesn't
the ECP is a statement of faith so there is little to "refute". when you've defined "economic calculation" to be trade, of course socialism can't do economic calculation. the entire point of the project is to get away from trade, to get away from the commodity fetish
>>2046
>implying there should be such a job as planner
do not obscure what are essentially political decisions behind some unelected experts. vidrel by zachariah and hardin seems to imply around 22:22 that the goals can be different in each geographical area, depending on political aims
>The only thing the ecp actually has any claim is exactly the time complexity of the calculation of a whole economy
I'd say the local knowledge problem is still relevant. but ancaps don't realize it cuts both ways. after all, who knows most about the process of production - the owners or the workers?


Unique IPs: 16

[Return][Go to top] [Catalog] | [Home][Post a Reply]
Delete Post [ ]
[ home / rules / faq ] [ overboard / sfw / alt ] [ leftypol / edu / labor / siberia / lgbt / latam / hobby / tech / games / anime / music / draw / AKM ] [ meta ] [ wiki / shop / tv / tiktok / twitter / patreon ] [ GET / ref / marx / booru ]