It is important to control the calcium or phosphorus concentrations in quarry products such as limestone. In some quarries in Japan, limestone blocks with low grade concentrations are blended with limestone blocks with high grade concentrations to satisfy the product quality requirement. In this study, some simulation methods were considered to develop the best mining plan for such quarries.
A quarry was divided into small rectangular blocks and a mining plan was considered as an optimal combination problem of these blocks. All possible combinations of limestone blocks were examined by a dynamic programming method, and almost the best solution was determined for a set of 600 blocks scale. However this method required large processing time for large-scale quarry; for example, it would be required 6,800 h for 2,400 blocks. On the other hand, genetic algorithm requires less than 6 h for 2,400 blocks, and it generates reasonable solutions whose quality rate is 80% of that generated by the dynamic programming method. Thus, to develop a long-term mining plan for large quarries, first genetic algorithm must be used to obtain approximate solutions, and then, the best solution should be determined by a dynamic programming method for short-term mining plan.
Moreover, the grade adjustment of the opposing limitations on calcium and phosphorus concentrations in limestone blocks, that is, the permitted calcium concentration has a lower limit whereas permitted phosphorus concentrations has an upper limit, was tried to solve. Finally, these methods can be easily applied to metal mining too.