Decision trees and sequence generators

Our approach in auto cash was to work with a container algebra, represented as a compact generator graph. The generator produced one of several typical pricing sequences to  match the sample.   Compact means optimally queued and we get node by node algebra, within some bound error.

Then we moved to decision trees for smart contracts, and consider a protocol as the simultaneous convolution of multiple decision trees.  From that model we considered decomposition, removing the tax and regulation decisions out to its own tree, for exaple.

Now we consider coherent decision trees that are within an isomorphism of each other, but differ in quant scaling. In particular, consider the decision tree for every tourist traveler planning a trip.  We consider if all trees are similar, but differ in size and quality. Middle class gravel vs high class travel; both have similar decision trees.

Do we get a index on decision trees? Yes,  and we thus get a starting algebra for general semantic networks.  We are going for closed and self generating decomposition and recomposition of generic semantic networks. The principle is to get probability values for decision points, then in a conserved resource situation we have a bound, the total probability. Singularity 2.0 under construction, it is about finding our Hamiltonians, the laws of economics. Up through the autocash layer, the assumption was agents who sphere pack with price, nothing more.