Introduction to Genetic Algorithms

10.9: Very Large Scale Integration (VLSI)

10.9 Very Large Scale Integration (VLSI)

10.9.1 Development of a Genetic Algorithm Technique for VLSI Testing

The objective of VLSI testing is to generate compact set of test vectors that has high coverage of manufacturing defects.

  • Stuck at fault modeling is the widely used fault modeling method in VLSI Testing.

  • Here nodes are assumed to be stuck at either "0" or "1", for the purpose fault modeling.

  • Testing methodology for a digital circuit is shown in Fig. 10.44.


    Fig. 10.44: Digital circuit

  • Test vectors are encoded as Binary bit stream.

  • Fitness function gives the number of faults covered by each test vector.

Consider the XOR circuit shown in the Fig. 10.45 below.


Fig. 10.45: XOR circuit

The description of given XOR Circuit is as follows:

  • Number of primary inputs is "2" and primary output is "1".

  • Each parent width is 2 bits.

  • In this example, XOR circuit has 12 fault sites and 24 stuck at faults.

  • For a fault free circuit, the output is "1" for a input vector [1,0].

  • If the circuit has a stuck at "0" at "a", the output response is "0". So input vector [1,0] detects stuck at 0 [SA0] fault at "a".

  • Like wise [1,0] can also detect SA0 fault at [a, c, d, g, h, z] and SA1 at [b, e, j].

  • Thus [1,0] can detect 9 out of a total of 24 faults and it's fitness is 0.375 [i.e. 9/24].

The experimental circuit is as given below in Fig. 10.46. Table...

UNLIMITED FREE
ACCESS
TO THE WORLD'S BEST IDEAS

SUBMIT
Already a GlobalSpec user? Log in.

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.

Customize Your GlobalSpec Experience

Category: Ground Fault Relays
Finish!
Privacy Policy

This is embarrasing...

An error occurred while processing the form. Please try again in a few minutes.