One major concerns in cellular manufacturing is to find how parts should be loaded and dispatched in a cell to enhance the advantages and diminish the disadvantages of group technology. This is known as group scheduling. Previous group scheduling researchers primarily focused on developing exhaustive and non-exhaustive heuristics in order to manage product flow through a manufacturing cell. In all the studies, the results considered a maximum of five machines in the cell and the processing time distribution assumption made was a third-order erlang distribution. The objective of this paper is to present new group-scheduling heuristics and compare them with existing heuristics for a seven-machine cell with three processing-time distribution assumptions. Computer simulation is used to examine the heuristics under eight experimental conditions. The results indicate that non-exhaustive heuristics improve on the exhaustive heuristics for average time in the system, average tardiness, and per cent tardy. The exhaustive heuristics improve on the non-exhaustive heuristics for set-up time and use. In particular, the DKFAM queue-selection heuristic is the best in all experimental conditions for many performance measures. The simulation model is coded in C++.