Large-scale simulation problems have extremely large CPU and memory requirements. Traditional methods for solving such problems involve parallel processing, which adds cost and complexity in programming. We propose a novel process based decomposition approach for distributed processing of such large-scale simulation problems. The proposed method splits up the large-scale problem based on the physics or computational units of the problem and distributes the computational load over a LAN. This method allows the problem to be solved in parallel in a distributed framework in a computationally efficient manner. The greatest advantage of this approach is that this method allows performance enhancement at little extra cost by better utilizing existing computational resources (CPU cycles). Based on the complexity of the decomposition, we categorize large-scale simulation processes into three classes. We present a distributed processor (OPTIMA) that enables application programs to be run in a distributed framework. Examples that demonstrate the efficacy of this method are presented. Results show that running the simulations in parallel on a distributed platform speeds up the process considerably with small hardware or software costs.