NEPTUNE has built-in mathematical algorithms that enable the coordination of information among multiple optimizations that are running in parallel and that improve the system’s performance. This is very different from the “multidisciplinary” capability that other commercial optimization codes offer in the form of either sequential optimization or multiobjective optimization. NEPTUNE allows for true system design through a network of optimizations. This approach leads to both system- and discipline-level improvements while meeting the designer’s specifications.
NEPTUNE has a graphical interface that enables users to set up the optimization trees graphically, built-in general-purpose interfaces that can exchange information with any solvers that are used for evaluating the objective functions and the constraints, and capabilities to construct metamodels (if desired) for replacing the high-fidelity solvers during the optimization computations. NEPTUNE also takes into account the uncertainty during the identification of the optimum configuration for a complex system, thus building in conservatism within the optimal solution and achieving both safety and optimality simultaneously.
The user defines the flowcharts for the system and for the discipline-level optimizations in NEPTUNE’s user interface. A data file is generated by the interface containing all the information that the user specified. This data file is used by NEPTUNE to automatically generate a MATLAB® script that contains:
The user runs the MATLAB optimization script that is created by NEPTUNE to complete the optimization analysis; results can be brought back to the NEPTUNE interface for postprocessing.
2890 Carpenter Rd Ste 1900
Ann Arbor, Michigan 48108