For the purpose of production scheduling, open-pit mines are discretized into three-dimensional arrays known as block models. Production scheduling consists of deciding which blocks should be extracted, when they should be extracted, and what to do with the blocks once they are extracted. Blocks that are close to the surface should be extracted first, and capacity constraints limit the production in each time period. Since the 1960s, it has been known that this problem can be cast as an integer programming model. However, the large size of some real instances (3-10 million blocks, 15-20 time periods) has made these models impractical for use in real planning applications, thus leading to the use of numerous heuristic methods. In this article we study a well-known integer programming formulation of the problem that we refer to as C-PIT. We propose a new decomposition method for solving the linear programming relaxation (LP) of C-PIT when there is a single capacity constraint per time period. This algorithm is based on exploiting the structure of the precedenceconstrained knapsack problem and runs in O mn log n in which n is the number of blocks and m a function of the precedence relationships in the mine. Our computations show that we can solve, in minutes, the LP relaxation of real-sized mine-planning applications with up to five million blocks and 20 time periods. Combining this with a quick rounding algorithm based on topological sorting, we obtain integer feasible solutions to the more general problem where multiple capacity constraints per time period are considered. Our implementation obtains solutions within 6% of optimality in seconds. A second heuristic step, based on local search, allows us to find solutions within 3% in one hour on all instances considered. For most instances, we obtain solutions within 1-2% of optimality if we let this heuristic run longer. Previous methods have been able to tackle only instances with up to 150,000 blocks and 15 time periods.

No abstract

In this work we consider the shortest path problem with uncertainty in arc lengths and convex risk measure objective. We explore efficient implementations of sample average approximation (SAA) methods to solve shortest path problems when the conditional value at risk and entropic risk measures are used and there is correlation present in the uncertain arc lengths. Our work explores the use of different decomposition techniques to achieve an efficient implementation of SAA methods for these nonlinear convex integer optimization problems. A computational study shows the effect of geometry, uncertainty correlation and variance, and risk measure parameters on efficiency and accuracy of the methods developed. Data and the online supplement are available at https://doi.org/10.1287/ijoc.2017.0795 .

Solving large scale nonlinear optimization problems requires either significant computing resources or the development of specialized algorithms. For Linear Programming (LP) problems, decomposition methods can take advantage of problem structure, gradually constructing the full problem by generating variables or constraints. We first present a direct adaptation of the Column Generation (CG) methodology for nonlinear optimization problems, such that when optimizing over a structured set X plus a moderate number of complicating constraints, we solve a succession of 1) restricted master problems on a smaller set S ⊂ X and 2) pricing problems that are Lagrangean relaxations wrt the complicating constraints. The former provides feasible solutions and feeds dual information to the latter. In turn, the pricing problem identifies a variable of interest that is then taken into account into an updated subset S ⊂ X .Our approach is valid whenever the master problem has zero Lagrangean duality gap wrt to the complicating constraints, and not only when S is the convex hull of the generated variables as in CG for LPs, but also with a variety of subsets such as the conic hull, the linear span, and a special variable aggregation set. We discuss how the structure of S and its update mechanism influence the number of iterations required to reach near-optimality and the difficulty of solving the restricted master problems, and present linearized schemes that alleviate the computational burden of solving the pricing problem.We test our methods on synthetic portfolio optimization instances with up to 5 million variables including nonlinear objective functions and second order cone constraints. We show that some CGs with linearized pricing are 2-3 times faster than solving the complete problem directly and are able to provide solutions within 1% of optimality in 6 hours for the larger instances, whereas solving the complete problem runs out of memory.