About Decision Rules
Many complex decision making problems can be reduced to a set of IF...THEN...ELSE rules.
This is especially true in making decisions in the context of operations
tracking and process flow management. Some examples include:
- IF the part being scanned into a manufacturing operation is not on
the bill of material (BOM) for the route operation step THEN warn the
operator and prevent them from recording the part in.
- IF the available inventory of a part falls below its safety stock
level THEN send an Alert message to the Purchasing Manager
- IF the data capture station has a weighing scale attached THEN use
weight from weighing scale ELSE have user manually input the weight.
There are several things to note here:
- All the rules are simple in format but almost all require complex
extraction of data from a database.
- Some are immediate, such as to prevent or warn an operator, others
need to be carried out periodically.
- For point-of-action data capture, there are a limited set of rules,
which may or may not be applicable in each circumstance.
- Rules periodically applied may have greater variability in form and
content than point-of-action rules.
For this reason:
- Rules related to data capture are expressed in the form of settings
in Excel spreadsheets, which can be imported and used to set which
rules are to be applied and their parameters. Alternately, for commonly
used rules, these settings can be manually entered through screens, such
as the Systems Administrator screens.
- Rules related to user defined parameters to be captured at
point-of-data-entry are expressed as Excel spreadsheets which are
imported and stored in a database.
- Rules related to data formats and their interpretation are entered
as HLDO spreadsheets and translated to XML metadata for use at run-time.
- Rules related to periodic evaluation of data are expressed as Python
scripts which can be imported and used by intelligent agent processes.
It might seem a little strange that a system, which is dedicated to
replacing Excel spreadsheets for manual data entry with automated data
collection, makes such extensive use of Excel spreadsheets for rules setup.
But Excel spreadsheets are familiar to most people and rule setup only has
to be done once, whereas data entry has to be done continuously on every
shift. They also provide an easy way for systems administrators, business
analysts, manufacturing engineers, and IT people to configure these systems
without having to learn a new paradigm.
The most critical things about the use of rules, as a form of Artificial
Intelligence, for decision making are:
- They are deterministic and do exactly what they are configured to
do.
- They require very little computing power and can therefore run in
real-time on small computers.
This is unlike using regenerative AI algorithms for decision making
where:
- These are based on self-adaptive statistical matrix correlators
which are prone to making false recommendations (hallucinations) -
remember there are lies, damn lies, and then, worst of all, statistics.
- These require execution using huge, expensive data centers in the
Cloud, which are slow, very expensive, and subject to the availability
of the Internet
For more details about the application of these rules, please see the
BellHawk User Manuals
Please click here to return to information
about the Tau Adaptor.
Copyright © SmartOpsMgt LLC 2026
