Parallel Genetic Algorithms (PGA) have been implemented in the past largely on parallel computers, and more recently on serial PCs. PGAs have been used successfully in solving many difficult optimization tasks. To gain further insight into the state and progress of the algorithm, we often need to extract useful information from the large amount of data generated from a PGA run, but this can be a difficult task. Many of the current PGA implementations often have no capability of visualizing an evolving GA population dynamically during execution time. In this paper, we describe an implementation of a finegrained parallel GA using Swarm, a multi-agent simulation tool originally developed at the Santa Fe institute. The PGA model developed is capable of visualizing dynamically the performance of an evolving GA population with plotted graphs on model parameter values in real time. This implementation also allows modification of some model parameter values during an optimization run, therefore offers advantages over many existing PGA implementations. We demonstrate the usefulness of the visualization techniques used in this PGA implementation using two optimization examples.