[OPE-L:7031] [OPE-L:526] Re: Re: Time subscripts and continuous time
clyder@gn.apc.org
Fri, 26 Feb 1999 09:42:14 +0000
>But, I'm not sure if anyone says that time isn't at all important
>in Marx. Even if one objects to the use of period analysis, as
>Paul C has done previously, there is no question that Marx at various
>times used a unit of time called a period. Indeed, it seems to me
>that the whole of VII would be incomprehensible without some concept
>of time.
>
I am not opposed to the use of the use of discrete timesteps in
the construction of digital simulation models of the economy, since
this is something that one is forced to do by the nature of the
computational technology. However, analytically the way to proceed
in modeling dynamical systems is to define a set of differential
equations in continuous time and use integration of these to derive
the time path of the system. As a later step, one can use discrete
timesteps in this integration if one is forced to use a digital
device in the modeling.
The continuous time model however should be conceptually prior, since
it is only computational convenience that forces discrete time steps
on us. If one builds, or has available to one an analogue computing
device, one can foregoe discretisation altogether. Phillips and Lange
in the 50s and 60s both experimented with analogue computing devices
for dynamical modeling of economies.
>
Paul Cockshott