Optimizing Personal Computer Configurationswith Heuristic-Based Search Methods

摘要

Given the diversity and limitedcompatibility for personal computer hardware,obtaining an (sub-)optimal configuration fordifferent usage restricted to some budgetlimits and other possible criteria can bechallenging. In this paper, we firstlyformulated these common configuration problemsas discrete optimization problems to flexiblyadd in or modify users' requirements. Moreinterestingly, we proposed two intelligentoptimizers: a simple-yet-powerful beam searchmethod and a min-conflict heuristic-basedmicro-genetic algorithm (MGA) to solve thisreal-life optimization problem. Theheuristic-based MGA consistently outperformedthe beam search and branch-and-bound method inmost test cases. Furthermore, our work opens upexciting directions for investigation.