Template realization of generalized branch and bound algorithm

Abstract

In this work we consider a template for implementation of parallel branch and bound algorithms. The main aim of this package to ease implementation of covering and combinatorial optimization methods for global optimization. Standard parts of global optimization algorithms are implemented in the package and only method specific rules should be implemented by the user. The parallelization part of the tool is described in details. Results of computational experiments are presented and discussed.

Straipsnyje pristatyta apibendrinto šaku ir režiu algoritmo šablono realizacija. Irankis skirtas palengvinti nuosekliuju ir lygiagrečiuju optimizacijos uždaviniu programu kūrima. Nuo uždavinio nepriklausančios algoritmo dalys yra idiegtos šablone ir vartotojui reikia sukurti tik nuo uždavinio priklausančiu daliu realizacija. Šablone idiegti keli lygiagretieji algoritmai, paremti tyrimo srities padalinimu tarp procesoriu. Pateikiami skaičiavimo eksperimentu rezultatai.

Key words:

• Templates
• parallel algorithms
• branch and bound algorithms
• combinatorial optimization
• global optimization