US 12,447,617 B1
Random key optimization for generating sequencing plans
Martin Schuetz, Brooklyn, NY (US); John Kyle Brubaker, San Diego, CA (US); Mauricio Resende, Seattle, WA (US); and Helmut Gottfried Katzgraber, Kirkland, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Dec. 15, 2022, as Appl. No. 18/066,983.
Int. Cl. B25J 9/16 (2006.01); G06N 10/80 (2022.01)
CPC B25J 9/1664 (2013.01) [G06N 10/80 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system, comprising:
at least one processor; and
a memory, storing program instructions that when executed by the at least one processor, cause the at least one processor to:
receive, via an interface, an optimization problem to generate a robotic motion plan, wherein the optimization problem is described according to a plurality of nodes that represent decision variables of the optimization problem;
determine and execute an optimization pipeline compatible with both classical computing hardware and quantum computing hardware to generate the robotic motion plan, wherein the optimization pipeline comprises:
application of a random key optimization technique using the classical computing hardware or the quantum computing hardware to the plurality of nodes to determine a vector of respective values for the plurality of nodes; and
decode of the vector of respective values for the plurality of nodes according to a motion planning constraint for the motion plan to generate the motion plan and determine a fitness score for the robotic motion plan; and
return, via the interface, the robotic motion plan generated by the determined optimization pipeline, wherein the robotic motion plan is one of a plurality of robotic motion plans generated according to the random key optimization technique, the robotic motion plan being selected according to the fitness score for the robotic motion plan.