D.C. programming approach for solving an applied ore-processing problem

Abstract

This paper was motivated by a practical optimization problem formulated at the Erdenet Mining Corporation (Mongolia). By solving an identification problem for a chosen design of experiment we developed a quadratic model that quite adequately represents the experimental data. The problem obtained turned out to be the indefinite quadratic program, which we solved by applying the global search theory for a d.c. programming developed by A.S. Strekalovsky [13]-[15]. According to this d.c. optimization theory, we performed a local search that takes into account the structure of the problem in question, and constructed procedures of escaping critical points provided by the local search. The algorithms proposed for d.c. programming were verified using a set of test problems as well as a copper content maximization problem arising at the mining factory.

Application of survival theory in Mining industry

“…According to Mendez et al [16], all known optimization approaches applied to flotation process can be classified into four general groups: mathematical techniques based methods without integer variables, mathematical techniques based methods with integer variables, heuristic methods, and genetic algorithms based evolutionary methods. There are many works devoted to optimization of ore-processing and flotation processes [10,12,15,17,7,18]. Maximization of concentrate grade formulated as a global optimization problem was considered in [10].…”

“…There are many works devoted to optimization of ore-processing and flotation processes [10,12,15,17,7,18]. Maximization of concentrate grade formulated as a global optimization problem was considered in [10]. The optimization problem was maximizing a quadratic function with a indefinite matrix over the box constraints.…”