对传统意义下的厂级负荷优化分配模型进行修正，同时考虑全厂供电煤耗率、污染排全放和负荷调整时间3个目标，提出厂级负荷分配的多目标优化模型。将多目标优化方法和多属性决策结合使用，研究多目标优化指导的问题。针对非劣分层遗传算法(nondominated sorting genetic algorithm II，NSGA-II)易于局部收敛的特点，提出了并行的NSGA-II多目标优化结构，增加了Pareto前沿的多样性，为决策提供丰富的信息。引入基于基点和熵的多属性决策方法，对Pareto解集进行排序，得出最优解。对某火电厂进行实例分析，结果表明该方法能准确快速地完成多目标负荷分配优化，并给出正确的指导，具有一定的实用性。

The traditional model of load dispatch was modified, three objectives which include power coal consumption rate of the whole plant, pollution emission and load adjusting time were considered simultaneously and a multi-objective plant load dispatch model was established. The multi-objective optimization and decision guidance problem was studied through a hybrid approach which was combined multi-objective optimization algorithm with multiple attributes decision making method. According to the local convergence of nondominated sorting genetic algorithms-II (NSGA-II), a new parallel NSGA-II multi-objective optimization structure was proposed. The diversity of Pareto front was increased and more abundant information for decision-making was provided. The Pareto solution set was sorted and the optimal solution was obtained by introducing method of multi-object decision- making based on maximum deviations and entropy. A case study of a power plant was carried out. The result shows that this method is practical and it can accomplish multi-objective plant load dispatch accurately and quickly. Finally the correct guidance is provided.