There has been a great increase in the number of high-power electrical appliances on the smart grid. Together with the popularization of smart terminals, and the increasing power consumption from the demand-side, this has brought the difficulties of power consumption to consumers. In this paper, the demand side scheduling scenario is considered from the three aspects of distributed generation, utility power and residential power consumption. Their time-sharing price models are constructed. Then, we introduce three functions to measure dispatching performance: resident comfort, electricity consumption economy and load variance. We also construct a weighted optimization objective model based on the dispatching performance function. Given that a complex multi-party time-sharing electricity price model participates in the dispatching, we propose an improved genetic algorithm to dispatch electricity consumption of demand side to minimize the objective function. Here additional elite selection strategies and evolutionary reversal operations are added. This can effectively reduce the iteration time and find an optimal value. Then, the convergence of the proposed algorithm is proved theoretically. Finally, the effectiveness of the algorithm is verified by simulation, and the power consumption cost is reduced by 31.29% while meeting the comfort of the resident power consumption. This work is supported by the National Natural Science Foundation of China (No. 62103070).