| A Weighted Coding in a Genetic Algorithm for the Degree-Constrained Minimum Spanning Tree Problem |

| | On a set of hard graphs whose unconstrained **minimum** **spanning** **trees** are of high degree, a steady-state GA that uses the weighted coding identifies degree-constrained **spanning** **trees** that are on average shorter than those found by several competing algorithms. |

| | Patterns of symbols do not represent consistent substructures of **spanning** **trees**, so that crossover may generate offspring whose **trees** do not resemble the **trees** of their parents, and the mutation of even one symbol may change the represented **tree** radically [23, 30]. |

| | The weighted coding of **spanning** **trees** was implemented in an otherwise conventional steady-state GA. The algorithm selects chromosomes to be parents in tournaments of size three, and generates offspring from them via uniform crossover and a mutation that resets each gene to a new random value with a small probability (position-by-position mutation). |

| www.acm.org /conferences/sac/sac2000/Proceed/FinalPapers/EC-11/www_wdmst.html (0 words) |