Casting Rubik’s Group into a Unitary Representation for Reinforcement Learning

Abstract Rubik’s Cube is one of the most famous combinatorial puzzles involving nearly 4.3 × 10 19 possible configurations. However, only a single configuration matches solved one. Its mathematical description expressed by group, whose elements define how its layers rotate. We develop unitary representation group and quantum formalism to describe based on geometrical constraints. Using particle states, we cubies as bosons for corners fermions edges. By introducing set four Ising-like Hamiltonians, managed global ground state all Hamiltonians. To reach Hamiltonian operators, made use Deep Reinforcement Learning algorithm reward. The successfully through phases, each phase driven corresponding reward energy spectrum. call our QUBE , it employs mechanics tackle problem solving Cube. Embedding problems into suggests new algorithms future implementations hardware.