The Test: Benchmarking the Best Nest

Benchmarking against a current method of producing parts using real world requirements is the only efficacious way to determine if new methods, such as Vision Emulation and Multi-Dimensional-Combinatorial-Nesting technology, are indeed improvements. Always having the highest level of automated productivity and cost cutting technology is quantifiable. Quantify a specific number of parts that are actually used in production; only then can optimized manufacturing metrics be compared and the best nest discovered.

Past methods of nesting have followed rules that reduce the very large number of combinations to a manageable few. The purpose of a heuristic rule set is to eliminate most of the calculations required to obtain a solution. Most nesting heuristics use some variation of the “First Fit” heuristic. The “First Fit” heuristic uses the following general steps:

1. Order the parts from largest to smallest.
2. Place the largest part on the raw material.
3. Rotate the part to the orientation that brings it closest to the nesting corner.
4. Place the next largest part on the raw material.
5. Rotate the part to the orientation that brings it closest to the nesting corner.
6. Repeat steps 4 and 5 until all parts are nested or until no more parts will fit.

This method is modified by adding fewer or more rotations, placing some parts in holes of other parts, reordering the list by other attributes such as longest or greatest perimeter. According to Michael D Lundy P.E. and President of Blue Springs, Missouri-based, Optimation, “In all cases heuristics are blind. In order to prevent the heuristic from selecting a nest that is infeasible, constraints are applied to the list. A filter is applied to prevent parts that are due in the future from being considered. Another filter will limit the list of parts to insure that too many tools are not required by the set of parts in the list. These filters limit many of the possibilities that are feasible. The result is a sub-optimal nest that is feasible but wastes material and productivity.”

Lundy’s company approached this problem with a dynamic method of evaluating optimal results. Multi-Dimensional-Combinatorial-Nesting technology (MDCN™) is a new technology which transcends previous nesting methods. Previously most nesting algorithms were simple heuristics, and did not consider the alternatives; this new methodology guarantees the highest efficiency while insuring that production schedule and priorities are optimized. Lundy suggested, “Standard heuristics crash badly when presented with real world problems. MDCN looks at many different dimensions affecting cost. Schedule, hot parts, material efficiency, order completion, tool optimization, common cutting, torch load, and other costs make finding the optimal nest Multi-Dimensional.”

Unlike heuristics, the combinatorial nesting approach uses fathoming to eliminate infeasible solutions and sub-optimal solutions. MDCN converges, learning as it works to find the optimal solution. Each new solution becomes the benchmark to best when comparing alternative combinations of parts; any solution that is not better or can not improve on the best solution found, is not considered.

When Vision Emulation is applied with MDCN, the part selection process is greatly enhanced. By “looking” which parts will fit into holes or small open areas in the nest, companies can see which orientation would fit best. Vision Emulation works much like a human being; when a person nests parts they look at all of the parts and their individual shapes. It is obvious that some parts will fit together well and other parts will fit inside of holes. Currently most nesting processes require a large number of rotations in a trial and error approach to orientation; these antiquated nesting systems rotate parts in small increments and try to fit the part after each rotation. This is time consuming and regularly leads to wrong results. Some heuristic nesting solutions brag about rotating parts at one degree increments; a part may need to be rotated 124.372 degrees to fit optimally. Viewing larger parts helps fabricators and other manufacturers to select the best combinations of parts. Vision Emulation brings a powerful toolset for finding and placing the right parts in combination with other parts to achieve the optimal nest. Thomas R. Cutler is the President & CEO of Fort Lauderdale, Florida-based TR Cutler, Inc. (www.trcutlerinc.com). Cutler is the founder of the Manufacturing Media Consortium of three thousand journalists and editors writing about trends in manufacturing. Cutler is a member of the Society of Professional Journalist and author more than 300 feature articles annually regarding the manufacturing sector. Cutler can be contacted at trcutler@trcutlerinc.com..... See More Details.