Purpose

---

  The anneal class performs simulated annealing to maximize a fitness
  function. This is a little different than normal, where annealing
  minimizes a penalty (or energy) function.
  Simulated annealing maximizes a function by taking random steps in
  parameter space from the current input. At each step, the function
  to maximize is evaluated given the current input. If the input
  yields a higher fitness than the previous input, it is
  automatically accepted and the old input discarded. If it is lower,
  than it is accepted with a probability that decreases as the
  simulation continues. The process continues, drawing new random
  inputs and occasionally swapping out the old inputs, until the
  simulation ends. Thus, the process transitions from a random walk
  to a hill-climbing algorithm; this helps prevent getting stuck at
  local maxima.

Category

---

  Optimization

Modification History

---

  Feb 2010: Written by Chris Beaumont