Chris Fawcett

AClib: A Benchmark Library for Algorithm Configuration

Modern solvers for hard computational problems often expose parameters that permit customization for high performance on specific instance types. Since it is tedious and time-consuming to manually optimize such highly parameterized algorithms, recent work in the AI literature has developed automated approaches for this algorithm configuration problem [1, 3, 10, 11, 13, 16].

HAL: a framework for the automated analysis and design of high-performance algorithms

Sophisticated empirical methods drive the development of high-performance solvers for an increasing range of problems from industry and academia. However, automated tools implementing these methods are often difficult to develop and to use. We address this issue with two contributions. First, we develop a formal description of meta-algorithmic problems and use it as the basis for an automated algorithm analysis and design framework called the High-performance Algorithm Laboratory. Second, we describe HAL 1.0, an implementation of the core components of this framework that provides support for distributed execution, remote monitoring, data management, and analysis of results. We demonstrate our approach by using HAL 1.0 to conduct a sequence of increasingly complex analysis and design tasks on state-of-the-art solvers for SAT and mixed-integer programming problems.

Static and Dynamic Portfolio Methods for Optimal Planning: An Empirical Analysis

Combining the complementary strengths of several algorithms through portfolio approaches has been demonstrated to be effective in solving a wide range of AI problems. Notably, portfolio techniques have been prominently applied to suboptimal (satisficing) AI planning. Here, we consider the construction of sequential planner portfolios for domainindependent optimal planning. Specifically, we introduce four techniques (three of which are dynamic) for per-instance planner schedule generation using problem instance features, and investigate the usefulness of a range of static and dynamic techniques for combining planners. Our extensive empirical analysis demonstrates the benefits of using static and dynamic sequential portfolios for optimal planning, and provides insights on the most suitable conditions for their fruitful exploitation.

Portfolio Methods for Optimal Planning: An Empirical Analysis

Combining the complementary strengths of several algorithms through portfolio approaches has been demonstrated to be effective in solving a wide range of AI problems. Notably, portfolio techniques have been prominently applied to suboptimal (satisficing) AI planning. Here, we consider the construction of sequential planner portfolios for (domain-independent) optimal planning. Specifically, we introduce four techniques (three of which are dynamic) for per-instance planner schedule generation using problem instance features, and investigate the usefulness of a range of static and dynamic techniques for combining planners. Our extensive experimental analysis demonstrates the benefits of using static and dynamic sequential portfolios for optimal planning, and provides insights on the most suitable conditions for their fruitful exploitation.