This paper presents an algorithm that aids the controls engineer in specifying a sensor and actuator configuration for regulation of large-scale, linear, stochastic systems such as a large space structure model. The algorithm uses a linear quadratic Gaussian controller, an efficient weight-selection technique based on successive approximation, and a measure of sensor and actuator effectiveness to specify a final sensor and actuator configuration. This configuration enables the closed-loop system to meet output specifications with minimal input power. The algorithm involves no complex gradient calculations and is numerically tractable for large linear models, as demonstrated by the solar optical telescope example in this paper. Additionally, the algorithm provides the controls engineer with information on the important design issues of actuator sizing, reliability, redundancy, and optimal number. HE advent of the Space Shuttle makes the large space structure (LSS) an imminent reality. These future space structures will be measured in kilometers and, of necessity, will be lightweight and highly flexible (light damping). Standard LSS missions will include power generation, surveillance, astronomy, and communications. These missions will require stringent pointing accuracy, shape control, and vibration suppression. To satisfy these demanding mission requirements, the LSS will almost certainly require an active, regulator-type controller with multiple sensors and actuators located throughout the structure.i-6 Furthermore, given the size of an LSS, there will be a large set of admissible sensor and actuator locations. The controls engineer then faces the problem of selecting a limited number of sensor and actuator locations to best the LSS mission. The term best in this paper means achieving the LSS output specifications with minimal actuator power. The solution to this sensor and actuator selection (SAS) problem needs at least the following ingredients: 1) A specific closed-loop control law structure; 2) a technique for systematically adjusting (tuning) control law parameters to achieve output specifications with minimal power; and 3) a technique to evaluate the effectiveness of possible sensor and actuator configurations in achieving output specifications with minimal power. This paper incorporates the preceding ingredients into an algorithm that solves the SAS problem for an LSS modeled as a linear stochastic system. Some necessary background information on linear stochastic systems and linear quadratic Gaussian (LQG) control theory are presented in Sec. II along with a formal statement of the SAS problem. Section III contains two important selection theorems, four pertinent facts, and the general flow of the SAS algorithm. Results from the SAS algorithm applied to the controller design for a large space telescope are presented in Sec. IV, and Sec. V contains concluding remarks.