THE LOCAL SEARCH ALGORITHM

initialize X
while (fitness(X)<goodenough) {       // fitness<-representation
   X' = mutate(X)                     // operators Reproduction/Variation
   if (fitness(X')>fitness(X)) X=X'   // Selection
}

1. Restart
    As long as it doesn't devolve into random search.
    watch out for high numbers of local opt or deceptive landscapes

2. Simulated Annealing

   k = 0
   T = T_max     // starting temp (hot!)
   while (T>T_min) {
       for (i=0; i<interations(k)) {
           X' = mutate(X)
           if (fitness(X') > fitness(X)) {
              X = X'   // good move
           }
           else {      // is bad move OK?
              bad = fitness(X) - fitness(X')  // how bad is it? (positive)
              if ( prob(Exp[-bad/T]) ) X = X'
           }
       }
       T = coolit(T, k)   // T *= 0.99
       k++
   }