Simcon program; conflict with leaky integrators

[From Bill Powers (951014.1125 MDT)]

Kent McClelland --

I don't know if you have Simcon 4.5 running. Here is a Simcon program
for a conflict between two systems with leaky-integrator outputs. The
leaky integrators are Zocher "amplifiers" with a time-constant of 20
seconds. While the reference signals are different, the outputs are
large and opposing. When they are the same, the outputs come to the same
value that is a fraction of the conflicted value.

···

======================================================================
title CONFLICT WITH LEAKY INTEGRATORS
time 1000 1

# SETTING UP THE REFERENCE SIGNALS

# generate variable reference signal for both systems
rv generator puls 0 300 33

# constant component of reference signal
rc const 50

# effective reference signals, common plus or
# minus the pulsed reference signal to create conflict
# followed by no conflict

r1 summator rc 1 rv 1 # common plus pulsed
r2 summator rc 1 rv -1 # common minus pulsed

# SETTING UP THE ENVIRONMENTAL FEEDBACK CONSTANTS

k1 const 0.5
k2 const 0.5

# CONTROL SYSTEMS

e1 comparator r1 cv # error signals
e2 comparator r2 cv # controlled variable is same for
                       # both systems (causing conflict)

# output = leaky integral of error signal
o1 amplifier e1 0 20 20.0
o2 amplifier e2 0 20 20.0

# cv depends on contributions from entities 1 and 2
cv summator o1 1.0 o2 1.0

group cv e1 o1
group cv e2 o2

print cv o1 o2 r1 r2
plot 1 0.2 0.2 1 1

If you already know how to run Simcon, fine. If you don't, ask for
clarifications.

I've been trying to get Wolfgang's Simcon 5.0 to compile and run. Having
some problems, so we can use 4.5 for now.
-----------------------------------------------------------------------
Best,

Bill P.

[From Kent McClelland (951014.1530 CDT)]

Bill Powers (951014.1125 MDT)

Kent McClelland --

I don't know if you have Simcon 4.5 running. Here is a Simcon program
for a conflict between two systems with leaky-integrator outputs. The
leaky integrators are Zocher "amplifiers" with a time-constant of 20
seconds. While the reference signals are different, the outputs are
large and opposing. When they are the same, the outputs come to the same
value that is a fraction of the conflicted value.

Bill, this sounds quite interesting, but I've never picked up Simcon
because I do most of my work on a Mac. I'll retrieve Simcon from the WWW
and see if I can get it installed on a PC somewhere locally in order to try
your program out. Thanks for following through on my question.

By the way, could I get some source code from you for a standard tracking
program like the type Tom Bourbon uses in his experiments? I have a
student working for me this semester who says he can port a program like
that to the Mac if it's written in Basic or C (or preferably C++). Thanks.

Kent