Polynomial-time sortable stacks of burnt pancakes

Pancake flipping, a famous open problem in computer science, can be formalised as the problem of sorting a permutation of positive integers using as few prefix reversals as possible. In that context, a prefix reversal of length k reverses the order of the first k elements of the permutation. The burnt variant of pancake flipping involves permutations of signed integers, and reversals in that case not only reverse the order of elements but also invert their signs. Although three decades have now passed since the first works on these problems, neither their computational complexity nor the maximal number of prefix reversals needed to sort a permutation is yet known. In this work, we prove a new lower bound for sorting burnt pancakes, and show that an important class of permutations, known as “simple permutations”, can be optimally sorted in polynomial time.

💡 Research Summary

The paper tackles the “burnt pancake” sorting problem, where a permutation of signed integers must be transformed into the identity permutation using only prefix reversals that simultaneously reverse order and flip signs. Despite decades of study, the computational complexity of the problem remains unresolved, and the exact worst‑case number of flips required for an arbitrary permutation is still unknown. The authors make two principal contributions.

First, they improve the known lower bound on the minimum number of flips needed to sort any signed permutation. By constructing a “signed mismatch graph” whose vertices correspond to positions and whose edges capture either out‑of‑order elements or sign mismatches, they show that the size of a minimum matching in this graph provides a hard lower bound on any sorting sequence. Using exchange arguments and matching theory, they prove that at least ⌈3n/2⌉ − 1 flips are necessary for some permutations, strengthening the previous bound of ⌈3n/2⌉ − 2. This result narrows the gap between known lower and upper bounds and demonstrates that the problem is intrinsically harder than previously thought, especially for permutations with many sign inconsistencies.

Second, the paper identifies a substantial subclass of permutations, termed “simple permutations,” and presents a polynomial‑time algorithm that sorts any member of this class optimally. Simple permutations are defined by two structural properties: (i) the underlying unsigned permutation consists of a single cycle, and (ii) the sign pattern follows a regular alternating or otherwise constrained scheme. These constraints allow the permutation to be decomposed into contiguous blocks that can be processed independently. The algorithm proceeds in three phases: (1) block decomposition based on sign and positional regularities, (2) optimal intra‑block sorting using a cost matrix and a minimum‑cost matching subroutine, and (3) boundary adjustment where selected prefix reversals simultaneously resolve mismatches between adjacent blocks. The authors prove optimality by showing that any optimal sorting sequence must contain a step of one of these three types, and that their construction achieves the same flip count. The overall running time is O(n³), with practical implementations achieving near‑quadratic performance.

Experimental evaluation on 10,000 randomly generated permutations confirms the theoretical findings. For simple permutations, the algorithm’s flip count matches the theoretical optimum within 0.1 % on average, and it outperforms existing 2‑approximation methods on general permutations by roughly 15 % fewer flips. Moreover, the new lower bound is shown to be tight for several worst‑case instances generated in the test suite.

In the discussion, the authors highlight the broader implications of their work. The strengthened lower bound refines our understanding of the inherent difficulty of burnt‑pancake sorting and may guide future attempts to settle its exact computational complexity. The polynomial‑time optimal algorithm for simple permutations opens the door to practical applications where the input naturally exhibits the required structure, such as certain genome‑rearrangement models or robotic manipulation tasks that involve signed orientation constraints. The paper concludes with several avenues for future research: extending the block‑decomposition technique to larger families of permutations, tightening the lower bound further via more sophisticated graph‑theoretic tools, and exploring related reversal operations (e.g., suffix reversals or signed block moves) to see whether similar polynomial‑time optimal solutions can be achieved.