Craig Beebe

Learn More
An MCM's increased throughput and dense circuitry can easily result in failure if the board contains \hot spots". Therefore, an accurate thermal model of an MCM was needed in the development o f a new placement algorithm designed to consider both total net length and heat constraints. This algorithm uses a combination of simulated evolution and simulated(More)
Typically, placement algorithms attempt to minimize the total net length of a printed circuit board (PCB). However, an MCM's increased throughput and dense circuitry can easily result in failure if the board contains “hot spots”. Therefore, an accurate thermal model of an MCM was needed in the development of a new placement algorithm designed to(More)
This paper presents an activation scheme for use with Hopfield neural network algorithms that guarantees a valid solution for a particular category of problems. The technique monitors the appropriate neurons and heuristically controls their activation function. As a result it has been possible to eliminate several constraint terms from the energy function(More)
  • 1