| 摘要 |
There is disclosed herein a method comprising: (a) receiving first user input specifying two or more algorithms and specifying one or more connections between the two or more algorithms, wherein each of the connections is a connection between an output of one of the algorithms and the input of another of the algorithms; (b) receiving second user input specifying one or more data structures consistent with the two or more algorithms; (c) operating on the one or more data structures in response to third user input in order to build one or more models for the two or more algorithms, wherein the one or more models include one or more uncertainty variables and one or more decision variables; (d) receiving fourth user input specifying an optimizer; and (e) performing an optimization using the user-specified optimizer in order to optimize an objective function, wherein said performing the optimization includes: (e1) generating one or more values of the one or more decision variables respectively; (e2) invoking a plurality of executions of the two or more algorithms according to the one or more connections, wherein each of the executions operates on a corresponding set of one or more instantiated models determined by the one or more models, the one or more decision variable values and a corresponding set of one or more instantiated values of the one or more uncertainty variables; (e3) computing one or more statistics based on output of the two or more algorithms from said plurality of executions; (e4) computing a value of the objective function from the one or more statistics; and (e5) repeating (e1) through (e4) a number of times. |
