Gerver Sofa
A candidate optimal shape for the moving sofa problem proposed by Joseph Gerver in 1992, assembled from 18 line segments and circular arcs, with an area of approximately 2.2195 and symmetry about the corridor's diagonal.
Timeline (last 90 days)
The developer Abhilekh Meda successfully reconstructed Gerver's sofa shape using the PPO algorithm.
Gerver's sofa consists of 18 curve segments, including straight segments and circular arc segments, and is overall symmetric about the corridor's diagonal.
The sofa shape found by PPO is highly similar to Gerver's solution, indicating that it at least converged to a local solution close to the global optimum.
In 1992, Joseph Gerver constructed a better shape with an area of about 2.2195 and proved its optimality.
The shape generated after PPO training has an area close to the theoretical optimum of Gerver's sofa (about 2.2195), but the method is approximate and the training process may be unstable.
Gerver's proof of the sofa's optimality is based on a meticulous analysis of the constraint conditions at each contact point during the sofa's motion, using variational methods and optimal control theory to derive the boundary conditions satisfying corner-turning feasibility, ultimately proving that the area upper bound of his construction cannot be surpassed.
Gerver's proof of the sofa's optimality remains to this day the only widely accepted rigorous optimality argument on this problem.
All Facts (7)
In 1992, Joseph Gerver constructed a better shape with an area of about 2.2195 and proved its optimality.
60%UnverifiedThe sofa shape found by PPO is highly similar to Gerver's solution, indicating that it at least converged to a local solution close to the global optimum.
50%UnverifiedGerver's sofa consists of 18 curve segments, including straight segments and circular arc segments, and is overall symmetric about the corridor's diagonal.
50%UnverifiedThe shape generated after PPO training has an area close to the theoretical optimum of Gerver's sofa (about 2.2195), but the method is approximate and the training process may be unstable.
50%UnverifiedGerver's proof of the sofa's optimality is based on a meticulous analysis of the constraint conditions at each contact point during the sofa's motion, using variational methods and optimal control theory to derive the boundary conditions satisfying corner-turning feasibility, ultimately proving that the area upper bound of his construction cannot be surpassed.
50%UnverifiedGerver's proof of the sofa's optimality remains to this day the only widely accepted rigorous optimality argument on this problem.
50%UnverifiedThe developer Abhilekh Meda successfully reconstructed Gerver's sofa shape using the PPO algorithm.
50%