Dynamic economic dispatch (DED) is an important problem in power system generation, operation, planning, and control. The objective of the DED problem is to schedule power generation for the online units over a time horizon, satisfying the unit and ramp rate constraints. Here, valve point loading effects that cause nonsmoothness of the objective function is also considered while solving the DED. The accuracy of the solution not only depends on the optimal scheduling of generating units, but it also lies in accuracy while estimating transmission system losses. Generally, B-loss coefficients are used in estimating transmission losses. However, in the literature, A-loss coefficients are found to be at par with B-loss coefficients in estimating transmission system losses. Therefore, in this paper, the performance in estimating the transmission system losses using A-loss coefficients are investigated through the solution of the DED problem. Here, a recently evolved heuristic search technique called the gravitational search algorithm is used for solving the DED. The feasibility of the proposed method is tested and validated on standard benchmark test systems such as the IEEE 30-bus system, IEEE 39-bus system, and IEEE 118-bus system. All simulations are carried out using SCILAB 5.4 (www.scilab.org), which is open-source software.