When a system is restored after an extended outage, the demand is greater than that was before the outage. The increase in the demand is due to the cold load pickup (CLPU) condition. The attempt to restore such load simultaneously will cause an excessive loading on the network elements in addition to limits violation. For these reasons, step-by-step restoration was suggested to overcome such problem. However, this way of restoration requires a long time for the complete system to be back to service. In this paper, a particle swarm optimization (PSO) algorithm is introduced to find the optimal sizing and siting of distributed generation (DG) units for simultaneous single step restoration of transmission and distribution networks. The aim of the proposed DG allocation algorithm is to reduce the distribution transformer loss of life due to overloading and to reduce the lost load and hence the additional power demand caused by CLPU. The capacity of the DG required is determined on the basis of additional power demand and the load diversity preserved. The proposed algorithm is applied to the Egyptian 66 kV transmission network in the city of Alexandria including a 33-bus and a 69-bus, 11 kV primary distribution feeders.