FPGA simulator igre gomoku

PETEH, MAKS (author), Šter, Branko (mentor)

Drevesno preiskovanje Monte Carlo (Monte Carlo tree search, MCTS) je preiskovalna metoda, ki združuje natančnost drevesnega preiskovanja z naključnim vzorčenjem. Pri drevesno preiskovalni Monte Carlo metodi je treba izvesti veliko število simulacij v čim krajšem času. Ideja je hitro izvajanje velikega števila iger, ki bi na FPGA tekla hitreje kot na računalniškem procesorju. Cilj je razviti simulacijo igre Gomoku z igralcema, ki naključno izbirata poteze. V okviru diplomskega dela je bil razvit simulator, ki izbira poteze igralcev, jih zapiše v pomnilnik ter preveri zmagovalca. Za implementacijo je bil uporabljen jezik za opisovanje strojne opreme VHDL in XILINX urejevalnik, ter vmesnik za testiranje.

• Language: Slovenian
• Keywords: VHDL, Gomoku, Artix 7
• Work type: Bachelor thesis/paper
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2016
• PID: 20.500.12556/RUL-81054
• Publication date in RUL: 25.03.2016

Secondary language

• Language: English
• Title: Gomoku game FPGA simulator

Abstract

Monte Carlo tree search is a search method that combines the precision of tree search with random sampling. In the method Monte Carlo tree search is to be carried out a large number of simulations in the shortest possible time. The idea is to implement a large number of games that would FPGA run faster than on a computer processor. The aim was to develop a simulation game Gomoku with players who randomly choose moves. In the context of the thesis has been developed simulator that chooses pleyers moves which are written in the memory and verification of the winner. For the implementation was used the language to describe hardware, VHDL and XILINX editor and interface for testing.

• Keywords: VHDL, Gomoku, Artix 7