How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Fast multipole expansions and error — Examples
1 · Prerequisites
2 · Summary
Two separated planar charge clusters checks an order-one M2L polynomial and both certified 2D tails for two separated charges. Near/far split with exact near evaluation and truncated far error adds a same-leaf near charge and verifies that direct near evaluation contributes no truncation error. A clustered point set makes tree depth enter the work fixes two points and increases their separation depth to show why tree size and traversal counts must appear in the work hypotheses.
3 · Logical flowchart
4 · Definitions, theorems and proofs
None yet.
5 · Examples, counterexamples and false statements
A clustered point set makes tree depth enter the work
Example
This example shows why the operation theorem records tree depth, node count, and list counts explicitly. In the square , place two distinct unit charges at and for an integer , and refine a dyadic quadtree until every leaf has occupancy at most .
Facts & Assumptions
Given: The dyadic square partition, the two distinct points above, leaf occupancy , and the per-node passes of the planar FMM.
Finite planar quadtree and separated interaction lists: dyadic bisection halves every cell side length, and the level-restricted linear-size hypothesis explicitly bounds the number of nonempty cells by a constant times .
Fixed-order planar Laplace fast multipole procedure: the upward and downward passes visit every nonempty leaf and every parent-child edge whose two cells are nonempty.
Planar FMM operation count under explicit linear-size tree and list bounds: the conclusion requires .
Verification
At depth the cell along the corner path containing has side length . For every , , so both and lie in that same cell; at the next split the second point is on the midpoint boundary and is assigned to the upper child, while is assigned to the lower child. Therefore the tree has a nonempty path through depth and at least nonempty cells, although .
With fixed , is constant as varies, whereas [F2] makes the passes visit at least the nonempty parent-child edges on the path from step 1.1. This family violates the uniform node bound required in [F3], so the theorem's hypotheses do not give a uniform bound over arbitrarily clustered inputs.
Two separated planar charge clusters
Example
Use the dyadic root square . Put and ; the other two root children are empty. Within name its lower-left and lower-right children and . Within name its lower-left, lower-right, and upper-left children , , and ; its upper-right child is empty. Let the leaf threshold be and split each cell with more than two labels. Put unit charges at in , where . Put zero-charge labels at in , in , in , and in . Thus and have three labels each and split; their named children are leaves. In the recursive ordered cell pair , splitting the source exposes , and then splitting exposes the separated pair . The far difference disk has center and radius , so use its analytic logarithm real at . Take order .
Facts & Assumptions
Given: The six labels and their charges/positions above, the recursive cell-pair list rule, and the order-one planar FMM.
Finite planar quadtree and separated interaction lists: a non-admissible pair is split by the specified cell rule, a separated pair is recorded as far, and direct near pairs are recorded at leaves.
Fixed-order planar Laplace fast multipole procedure: P2M moments are , M2L uses the displayed local coefficients, and near values/derivatives are added exactly.
Global planar FMM potential and derivative error: one far pair with ratio has potential and derivative bounds and at order .
Verification
The root has six labels, and its nonempty children each have three, so threshold makes internal while all named grandchildren are leaves. The centers of are , their radii are , and their center distance is , so is not separated. Its equal side lengths make the rule split the source ; child has center and radius . The pair is not separated because its center distance is . Splitting the larger cell exposes centered at ; now . Thus the list recursion records as a far pair, and the zero-charge label at is the target at the target-cell center.
The two unit source offsets give and . With , [F2] gives the local coefficients and . At the target center this interaction returns potential and derivative . Other source labels have zero charge, so they add no contribution to this target.
The two nonzero source-target differences are and , so the exact potential is and the exact derivative is . Comparing with step 1.2, the absolute errors are and ; the logarithmic inequality follows from for . Zero-charge labels add no contribution.
For the only far interaction with nonzero source charge, , , and . The [F3] bounds are and ; all other far source cells have . Any direct near pairs from the zero-charge labels contribute exactly zero, so the errors in step 2.1 satisfy the theorem's global sums.
Near/far split with exact near evaluation and truncated far error
Example
Use root ; its two nonempty children are and . Name the lower children of by and . Name the lower-left, lower-right, and upper-left children of by , , and ; its upper-right child and the two upper root children are empty. Let leaf threshold and split each cell with more than two labels. Put unit charges at in , where , a charge at in , where , and a zero-charge target label at . Put zero-charge labels at in , in , and in . Then has three labels and has four, so both split and their named grandchildren are leaves. For the target , the recursion records as far and the charge- pair in as direct near. Use the analytic logarithm on the far difference disk that is real at ; the near difference uses its real logarithm. Take order .
Facts & Assumptions
Given: The seven labels, charges, and positions above, the recursive cell-pair rule, and the order-one algorithm.
Finite planar quadtree and separated interaction lists: the recursive rule separates the far pair from direct near particle pairs in the leaves.
Fixed-order planar Laplace fast multipole procedure: exact direct near values are added to the far local polynomial and its derivative.
Global planar FMM potential and derivative error: the far interaction at ratio has potential and derivative bounds and for .
Verification
The root has seven labels, and its nonempty children have three and four labels, respectively; threshold therefore splits all three, while the named grandchildren are leaves. The centers of are , their radii are , and their center distance is , so is not separated. The tie rule splits source , exposing with center and radius . The pair is not separated since its center distance is , so the larger target cell is split. The child has center and radius , and ; hence is far. The distinct charge- source and zero-charge target at lie in the same leaf , so their ordered pair is direct near.
The far source moments are and , so M2L for gives and ; at the far approximation is . The direct near contribution of the charge- source is exactly , because ; direct contributions from zero-charge labels vanish.
The exact far terms are and . Adding the exact direct near values gives total . The algorithm returns , so the errors are and ; the strict logarithmic bound follows from for .
For the only far interaction with nonzero source charge, , , and , so [F3] gives bounds and . All other far source cells have zero charge, and every near pair was evaluated exactly, so the errors from step 2.1 satisfy the theorem's global sums.