Browse Prior Art Database

State Machine Verification Expert

IP.com Disclosure Number: IPCOM000120647D
Original Publication Date: 1991-May-01
Included in the Prior Art Database: 2005-Apr-02
Document File: 2 page(s) / 92K

Publishing Venue

IBM

Related People

Barrett, KL: AUTHOR [+4]

Abstract

An Expert System capable of generating any combination of functional test cases under users' direction to test a family of RISC System/6000* chip set logic design.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

State Machine Verification Expert

      An Expert System capable of generating any combination of
functional test cases under users' direction to test a family of RISC
System/6000* chip set logic design.

      The expert system was conceived for generating "all possible
scenarios" for a state-machine.  The state-machine is described in
the form of the knowledge-base entities, e.g., the input and output
signals are represented as units in the knowledge-base, and the
state-transition rules are represented as forward-chaining rules.
The expert system thus provides a natural representation of the
state-machine; yet it is not a blind emulation of the state-machine
as envisioned by a logic designer, because it uses some heuristic and
causal knowledge to "prune" the search space of the inputs and the
states.  Thus, it is possible for the expert system to test
effectively all the combinations of the input signals and the
possible states of the state-machine in a "reasonable" time.  The
practical application of this system has been successfully achieved
for a sample of five state-machines that are representative of the
logic design of Application Specific Integrated Circuits (ASICs) used
in RISC System/60000 workstations.

      The conventional programming approaches soon grow too complex
to be manageable in the case of large state machines.  Since the
problem of combinatorial explosion is the same for both the
conventional and the expert systems approach, it is wiser to take
advantage of the search strategies available in the field of
artificial intelligence.  The simplest AI approach to this problem
appeared to be that of a rule-based expert system with
forward-chaining rules that allow an easy imbedding of the original
system (the state-machine) into the solution.

      The knowledge-based approach has the advantage that it can
generate explanations for the behavior of the machine as time
evolves, and it can mask out the details at a low level of operation
of the machine to generate output only at a "higher level" if so
required.  Also, the expert system has the character of an
object-oriented approach because of the classification of the data
structures into hierarchies of units that represent an enti...