Shipboard power systems that service large-scale dynamic loads and electric propulsion can significantly improve their performance by adding controllable energy storage. These systems require power management controls that consider dynamics like generator ramp rates, along with energy storage capacity and location. In the early-stage design process, many alternative designs are considered, and each requires a unique controller. This paper describes a numerical optimization technique that establishes an upper bound performance criteria without manually designing controllers for each system. The solution is a best case performance that assumes perfect future knowledge of the time-varying load. While unrealistic in real-time, this technique yields a fair comparison between competing architectures without the variability of different control methods. To demonstrate the concept, a notional multi-bus power system architecture is evaluated on a representative set of operational duties to illustrate comparisons between system attributes like energy storage power and efficiency ratings. A design trade study shows that the success rate for a baseline ship can improve from under 60% to nearly 100% by increasing generator power by 10% and energy storage capacity by 100%. These automated architecture benchmarks fit into a broader total ship optimization process, or can be used in human-driven trade studies. INDEX TERMS Electric ship, energy storage, power flow, power system optimization, power management, pulse loads, power system control, ship design.