CaDS Seminar 2026 - Feb. 10

Dr. Sandipan Mohanty (SDL Biology)

Snakes in tight boxes: looking for optimal arrangements under tight constraints

Abstract:

A few years ago, we found that the D-Wave Advantage quantum annealer can be used to find the global energy minimum of lattice protein chains for chains with up to 64 beads. The quantum annealer out-competed classical simulated simulated annealing by maintaining a 100% success rate of finding the ground state for all systems with up to the 64 beads. We have since extended our approach to other interesting problems.

In this talk, I will present results of a study in which we examined the search for the ground state of self-avoiding chains in very tiny 3D lattices, such that the chain covers all lattice sites. The absence of free sites poses a serious challenge to classical explicit chain Monte Carlo methods. We found that the quantum annealer has no difficulty finding the ground state of such systems, taking the minimum available annealing time of 6s, for various 48-beadchains on a 3x4x4 lattice. For comparison and verification, we implemented an optimized C++ program to exhaustively enumerate every possible arrangement of the chain in the lattice. The computation takes about 8500 core hours on JUSUF, compared to 6 seconds on JUPSI. Interestingly, while normal chain-based classical Monte Carlo simulations fail completely under these constraints, classical simulated annealing successfully determined the ground state when we used it with the binary encoding of the problem developed for the quantum annealer.