This paper presents an efficient method for managing Monte Carlo simulation experiments to select the optimal circuit design from a set of candidates. Simulation is a useful tool for evaluating and comparing circuit designs since it measures the impact of component variability. However, its use in circuit design has traditionally been limited to problems with a small number of design candidates due to its large computational requirements. We outline a solution method that has been successfully used in other contexts to improve the efficiently of simulation-based optimization. The method works in an iterative fashion to intelligently allocate a limited computing budget across multiple design alternatives in order to maximize the probability of correct selection. We illustrate the method's potential benefits for circuit design problems through two simple examples. The examples confirm that the method yields significant savings in computational time, making simulation-based experiments a feasible option for larger circuit design problems.
|Original language||English (US)|
|Number of pages||7|
|Journal||Proceedings of the IEEE Annual Simulation Symposium|
|State||Published - Jan 1 2001|