A parallel composite genetic algorithm for mine scheduling

被引：0

作者：

Lindon, LF ^{[1
]}

Goforth, D ^{[1
]}

van Wageningen, A ^{[1
]}

Dunn, P ^{[1
]}

Cameron, C ^{[1
]}

Muldowney, D ^{[1
]}

机构：

[1] Laurentian Univ, MIRARCO, Sudbury, ON P3E 2C6, Canada

来源：

Proceedings of the Ninth IASTED International Conference on Artificial Intelligence and Soft Computing | 2005年

关键词：

genetic algorithm; parallel algorithm; village; scheduling; optimization; diversity;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Mine scheduling is a multi-objective highly constrained optimization problem. Often months are spent by mine planners to achieve one feasible ore production solution. In order to assist in the process and to present alternatives with a higher likelihood of optimality, a parallel genetic algorithm for long-term scheduling of underground mines is developed. For the mine scheduling problems considered, a 2-dimensional map, of stopes is given, along with the mineral properties of each. It is required to schedule the extraction sequence of the ore from the slopes to meet the mine's objectives. The appropriateness of a schedule is determined by applying a fitness function. The fitness function assesses how well the schedule meets objectives and satisfies given constraints. In practice, the scheduling problem is simplified in order to obtain a solution in given time bounds. By modularizing the problem and employing a parallel algorithm with minimal communication requirements, a higher quality mine schedule may be found in given time bounds.

引用

页码：245 / 250

页数：6