hu.birot.OTKit.runableExamples
Class Delta_shape
java.lang.Object
hu.birot.OTKit.runableExamples.Delta_shape
public class Delta_shape
- extends java.lang.Object
This examples creates a candidate set with three candidates (A, B and C) and two constraints:
noB and noC. The tableau looks as:
The topology is as follows: Candidate B has two neighbors, both A and C. But candidates A
and C are not neighbors of each other, their sole neighbor is B. Consequently, A is globally
optimal, while C is (another) local optimum.
Given this tableau and topology, SA-OT will return both A and C in 50% of the cases,
independently of the cooling schedule. Yet, simulated annealing with Harmonic Grammar
will prefer the global optimum A; the frequency of returning C will diminish, as the cooling
schedule becomes slower.
The code of this class demonstrates how to encode a Gen, a topology, constraints and
hierarchies, and how to run experiments with them subsequently.
Method Summary |
static void |
main(java.lang.String[] args)
|
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
Delta_shape
public Delta_shape()
main
public static void main(java.lang.String[] args)
- Parameters:
args
-