Peaches and economics

Am I allowed to just move pics around the web?

Anyway, the subject is the distribution of peaches and why that should be viewed as a Shannon information problem. I start by assuming that peaches, as they move from orchard to consumer, is a minimum spanning tree over some geography. It is a finite stage, directed graph

So, under what conditions and how can the research model peach distribution as an information channel?

First, the how. View the entire production of peaches as a series of transactions of the form rate*volume. When viewing the whole channel, each transaction from orchard down, all transactions should be equally surprising.

The why is because any predictable change in some type of transaction would be noticed as a low entropy event, a free lunch, and soon disappear. For example, if I notice that peach jam is over stocked two years after a warm Georgian winter, then I would play the peach jam futures.

So, in the smooth world of variances, cyles would not exist. The EMH would say that noticeable cycles on goods flow would soon be corrected by inventory flow adjustments. The yield curve, the inventory growth of the production line, as the distribution of rates, should be Gaussian, right?, because that is the maximum variance container.

So we say that the peach network is computing the maximum entropy yield curve estimate of a drunkards walk. We also say it is a Gaussian noise Shannon channel.
But to human agents in the process it is best view as a Huffman encoder. The reason for the last is information rate over time wants to be minimized, moving stuff is expensive. So, the lower the precision, the sparser the spanning tree, so total information rate, the rate of one quant of information per time, is lowest.

When peaches are combined with the rest of the economy, they use the same size semi-trailer as other fruits. The food processors making jam, re-distributors, and all the rest, share equipment standards, energy systems, government, and money with each other. But if the economy is at all efficient, even these should be workable with cross entropy information type analysis. Say, oil. If the aggregate economy was working with finite low entropy, then most of the "bit allocations" would be in energy distribution. So, total dimensional should still be small enough for analysis.

Distorted channels have quantum restrictions, there is a minimum shelf space for most fruits at the grocer. Or the total shelf space changes with the season. These limits reduce precision, but they limit the dimensionality of the problem. Generally what that means is that peach distribution is smooth within a certain range; but a shock, like a cold winter, has non-linear effects. These enforced low entropy restriction appear on the spanning tree as unbalances. These cause variance estimations to suffer from non-smooth distributions.