Tightening the mixed integer linear formulation for the piecewise linear approximation in general dimensions
This work addresses computational efficiency in optimization for data approximation, but it is incremental as it builds on existing MILP and DC representation methods.
This paper tackles the problem of tightening the mixed-integer linear programming formulation for continuous piecewise linear approximations in arbitrary dimensions, resulting in significant improvements in solution times through strategies like fixing variables and introducing constraints.
This paper addresses the problem of tightening the mixed-integer linear programming (MILP) formulation for continuous piecewise linear (CPWL) approximations of data sets in arbitrary dimensions. The MILP formulation leverages the difference-of-convex (DC) representation of CPWL functions. We introduce the concept of well-behaved CPWL interpolations and demonstrate that any CPWL interpolation of a data set has a well-behaved version. This result is critical to tighten the MILP problem. We present six different strategies to tighten the problem, which include fixing the values of some variables, introducing additional constraints, identifying small big-M parameter values and applying tighter variable bounds. These methods leverage key aspects of the DC representation and the inherent structure of well-behaved CPWL interpolations. Experimental results demonstrate that specific combinations of these tightening strategies lead to significant improvement in solution times, especially for tightening strategies that consider well-behaved CPWL solutions.