From: Paul Cockshott (wpc@DCS.GLA.AC.UK)
Date: Tue Aug 21 2007 - 08:39:32 EDT
You are obviously right about the data collection issue. What is significant About Jacek Gondzio's work is that it achieves optimisations over problems much Larger than one would need for basic economic planning. After speaking to him I realise that he gets such large numbers of variables because He is doing stochastic planning taking into account many different scenarios. He Starts out with a basic problem having hundreds of thousands or millions of variables, And then projects forward to a large number of possible futures in order to select The current plan that is most future-proof, it is the thousands of different futures That are considered that push the number of varibles so high. This future-proofing, as Far as I know, goes far beyond anything attempted in the socialist countries. -----Original Message----- From: OPE-L [mailto:OPE-L@SUS.CSUCHICO.EDU] On Behalf Of Ian Wright Sent: 20 August 2007 22:52 To: OPE-L@SUS.CSUCHICO.EDU Subject: Re: [OPE-L] A startling quotation from Engels > I will be visiting Dr Gondzio at Edinburgh University tomorrow to discuss the work he has > been doing on solving linear programming problems with one billion variables with > novel algorithms. So this area is experiencing very significant advances in the last > few years. I wonder if the problem is not one of computation but of representation? One billion variables implies one billion explicit definitions of distinct commodity types. It requires work to create such definitions (e.g., defining the thousands of different kinds of apples) and to categorise concrete instances of commodities. Plus, new commodity types will appear and old ones will die, requiring continual maintenance work to maintain the active list of commodity types that constitute the dimensions of the IO matrix. Maybe you have a creative solution to this problem, or accept the social cost. But perhaps a linear programming representation of an economy is not a good one, even for planning?
This archive was generated by hypermail 2.1.5 : Fri Aug 31 2007 - 00:00:10 EDT