中文摘要：

为获得更低的平均候梯时间和长候梯率，提出一种基于知识的自适应电梯群控制策略．该控制策略汇集了区域权重控制算法、电梯运行操作知识以及层站召唤再分配规则，基于自适应的长候梯时间阈值，对长候梯层站召唤执行再分配操作，凸现了电梯群控制策略对复杂电梯交通的自适应性．与经典THV算法、基于知识的区域权重控制算法、人工免疫动态优化算法比较，该方法能获得更低的平均候梯时间和长候梯率．同时，其自适应能力使得该控制策略更易于应用在实际电梯群控制系统中．

英文摘要：

To achieve less average waiting time and less long-wait percent in elevator group control system, an adaptive talented-based control policy was developed. This novel control policy integrated the area weight control algorithm, elevator operation strategy and the principle of hall call re-allocation, where hall calls with long-wait time would be reallocated according to the adaptive long-wait time threshold from the real elevator traffic. This major feature indicates that the new control policy is adaptive to complex elevator traffic. Compared to the THV algorithm, the area weight control algorithm with talent and the dynamic optimization based on artificial immune system, the proposed talented-based adaptive method could achieve shorter average waiting time, lower long-wait percent. Furthermore, it is easier to apply this new policy to real elevator systems due to its adaptability.