supply-demand balancing

[bhaugen]

In Telegram, @pospi wrote:

I had an interesting revelation talking to Michel Bawens last night which you might want to factor into your designs. I'm not sure of the feasability of pulling the data out but I expect it should be simple-ish and @fosterlynn can probably confirm how easy it might be.

Anyway, the basic idea would be to show related inventory items when viewing an agent's inventory or a process tree, so that users can gauge the market saturation for their items. This enables people to avoid over-production and make more informed decisions about their production processes; there's no sense making more shoes if there are already hundreds of shoes present in the network. Thoughts?

We've thought about this a lot. We think it most likely needs to be analyzed at a higher (more aggregated) level than an particular inventory item or production resource type. A shoe inventory item is likely to be of a particular style and size, while related items might be all shoes of the same size, or of the same style, or all shoes of any size and style. Or maybe even all footwear: shoes, boots, sandals, etc.

Same with food or almost any other item.

We've done some work on an aggregate level in Locecon, for example:

• http://locecon.org/nova-story/
• http://locecon.org/clusters/network/25/qty/fn/

[Edit]: The Locecon model is compatible with the OCP/ValueFlows models, as you can see from this old comparison.

What both the Nova Scotia projects and the Mutual Aid Networks want to do is integrate the operational level (like OCP) with the aggregate level (like Locecon), so you can zoom up and down and in and out.

This would also enable several different types of analysis, for example supply and demand balancing and input-output analysis, as well as the algorithms listed in the Value Flows overview.

This is also where taxonomies like those @bum2 is working on could come in. The taxonomies would provide the levels of aggregation. But which taxonomy to use will often depend on the community and cluster (in Locecon terms), for example, the Nova Scotia fish taxonomy is different from a food taxonomy which is different from the OCP taxonomy or the Mutual Aid Network taxonomy.

A decentralized or federated implementation of economic network software would enable large-scale economic analyses, as outlined in this VF issue: https://github.com/valueflows/valueflows/issues/135

In an email, @fosterlynn wrote:

There is a current discussion over in the Open Food Network about how to make their installations interoperable that revolves around the food taxonomies. One of the current opinions is that they will have to have an intermediate mapping of resource types [a common "pivot" vocabulary], it is too difficult for lots of people to agree, there is long local history, etc. This seems like a useful idea to me. There is also an idea of using faceted classification of existing resource types, and having the combination of facet values be the "primary key" (so to speak) across contexts.

[edit] Examples of facets:

• shoe style and size
• fish species and fishing gear (trawler, trap, longline)
• food species, variety, growing method (conventional, organic)

[Note: this issue intro will probably be edited]