Designing market-based control with a genetic algorithm

Semi-active devices are low-cost as well as small, and, by changing the properties of low-power intensive structures, the aims of control are accomplished. On the other hand, the limited control force, which can be applied to the structure for each damper, causes more dampers to be used in structures compared to larger and stronger control devices which are more costly. These dampers coupled with sensors and the structure themselves make a complex dynamic system which is best controlled by a decentralized method, such as Market-Based Control (MBC). In MBC, the actuators and the supply energy are modeled as the buyer and the seller, respectively, in the market place. To dene the demand function of the buyer and the supply function of the seller, some weighting constants have to be chosen. The performance of the MBC correlates with prudent selection of the weighting constants. In this study, a novel method for designing MBC using a Genetic Algorithm (GA) is presented. The MBC approach is applied to three linear structures having ve, ten and twenty oors, and the resulting solutions show the merits of the new methods for tuning MBC as opposed to solutions using a centralized Linear Quadratic Regulator (LQR).