This document summarizes a study on using a genetic algorithm with a mutation operator based on the sandpile model of self-organized criticality to solve dynamic optimization problems. The sandpile genetic algorithm (GGASM) uses a 2D lattice to represent the population, where grains are dropped on the lattice at a rate g. When slopes exceed a critical value, grains topple to neighboring cells in avalanches whose sizes follow a power law distribution. This distribution determines the bit-level mutation rates. Experiments tested GGASM on dynamic problems with varying severity, frequency of changes, base functions, and grain rates g. Results showed the mutation rate distribution adapts to problem dynamics and GGASM outperforms algorithms without sandpile