Learning Evaluation Functions to Improve Optimization by Local Search

189 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Learning Combinatorial Optimization Algorithms over Graphs (2017, methodology): “Boyan and Moore [7] use regression to learn good restart rules for local search algorithms.”
A unifying framework for computational reinforcement learning theory (2009, background): “…[Crites and Barto 1996], discrete optimization problems such as job-shop scheduling and Boolean satisfiability [Zhang and Dietterich 1995; Boyan and Moore 2000; Lagoudakis and Littman 2001], channel allocation [Singh and Bertsekas 1997], algorithm selection [Lagoudakis and Littman 2000],…”
Instance-Specific Algorithm Configuration (2014, methodology): “While searching for a local optima, STAGE learned an evaluation function to predict the performance of a local search algorithm.”
Transfer in Reinforcement Learning via Shared Features (2012, methodology): “The X-STAGE algorithm (Boyan and Moore, 2000) uses features common across a class of tasks to transfer learned evaluation functions that predict the performance of a local search algorithm applied to an optimization task.”
Learning an Approximation to Inductive Logic Programming Clause Evaluation (2004, methodology): “Though not in the domain of ILP, Boyan and Moore's STAGE algorithm [23] use quadratic regression to approximate search "trajectories.”