Code optimization in FORM

Kuipers, Jan; Ueda, Takahiro; Vermaseren, J. A. M.

doi:10.1016/j.cpc.2014.08.008

Cited by 79 publications

(79 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Since the band is still broad, multiple samples are required to approach the global minimum, regardless of C p . This is governed by the R parameter (van den Kuipers et al, 2013b), and is consequently a topic for future research.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Combining Simulated Annealing and Monte Carlo Tree Search for Expression Simplification

Ruijl

Vermaseren

Plaat

et al. 2014

Proceedings of the 6th International Conference on Agents and Artificial Intelligence

View full text Add to dashboard Cite

Abstract:In many applications of computer algebra large expressions must be simplified to make repeated numerical evaluations tractable. Previous works presented heuristically guided improvements, e.g., for Horner schemes. The remaining expression is then further reduced by common subexpression elimination. A recent approach successfully applied a relatively new algorithm, Monte Carlo Tree Search (MCTS) with UCT as the selection criterion, to find better variable orderings. Yet, this approach is fit for further improvements since it is sensitive to the so-called "exploration-exploitation" constant C p and the number of tree updates N. In this paper we propose a new selection criterion called Simulated Annealing UCT (SA-UCT) that has a dynamic explorationexploitation parameter, which decreases with the iteration number i and thus reduces the importance of exploration over time. First, we provide an intuitive explanation in terms of the exploration-exploitation behavior of the algorithm. Then, we test our algorithm on three large expressions of different origins. We observe that SA-UCT widens the interval of good initial values C p where best results are achieved. The improvement is large (more than a tenfold) and facilitates the selection of an appropriate C p .

show abstract

Section: Resultsmentioning

confidence: 99%

“…At the previous ICAART conference the sensitivity to C p and N was presented (van den Herik et al, 2013b). At CCIS/BNAIC the sensitivity to R was recognized Kuipers et al, 2013b). We believe that the practical applicability of the algorithm will improve as the sensitivity to these parameters is harmonized.…”

Section: Introductionmentioning

confidence: 99%

Combining Simulated Annealing and Monte Carlo Tree Search for Expression Simplification

Ruijl

Vermaseren

Plaat

et al. 2014

Proceedings of the 6th International Conference on Agents and Artificial Intelligence

View full text Add to dashboard Cite

show abstract

“…• The more sophisticated output optimizations of Form (procedures to compactify the expressions in combination with the built in output optimizations [16]) are almost a requirement.…”

Section: Pos(ll2014)075mentioning

confidence: 99%

Xtreme manipulations

Vermaseren¹

2014

Proceedings of Loops and Legs in Quantum Field Theory — PoS(LL2014)

View full text Add to dashboard Cite

“…In this case one might need to update the PATH and PYTHONPATH environment variables accordingly. The package needs Form-4 [29,30], which should be installed separately by the user, in order to compute the expansions which are needed and for the generation of the corresponding source code.…”

Section: Installationmentioning

confidence: 99%

“…In order to simplify the generation of the numerator expansions and the corresponding source code needed by Ninja as input, we distribute with the library a small Python package called NinjaNumGen which uses Form-4 [28][29][30] in order to compute the expansions and produce an optimized source code which can be used by the library.…”

Section: Introductionmentioning

confidence: 99%