---
title: 'K-value: Algorithm Benchmarks & Kaon Decays'
url: https://www.emergentmind.com/topics/k-value
type: topic
---

# K-value: Algorithm Benchmarks & Kaon Decays

Searching arXiv for the cited papers and closely related terminology.
arXiv search query: id:1502.07725 OR all:"k-Leaf Spanning Tree Problem Admits a Klam Value of 39"
In the arXiv literature under consideration, the label associated with “K-value” appears in two distinct technical senses. In parameterized complexity, it refers to the **klam value** of an algorithm for the $k$-Leaf Spanning Tree problem, namely the largest integer $k_0$ such that the parameter dependence $f(k)$ in a running time $O^*(f(k))$ satisfies $f(k)<10^{20}$ for all $0\le k\le k_0$ [1502.07725]. In hadronic phenomenology, the symbols $F_K$ and $F_K'$ denote the weak decay coefficients of the kaon and of the first radial excitation $K(1460)$ in the extended NJL model [1906.06680]. The two usages belong to different domains and should not be conflated.

## 1. Scope of the term in different research contexts

The parameterized-complexity usage arises from the $k$-Leaf Spanning Tree ($k$-LST) problem, which asks whether an undirected graph $G$ contains a spanning tree with at least $k$ leaves. Here $k$ is the parameter, and the central quantitative object is the klam value attached to a parameterized running time [1502.07725].

The hadronic-physics usage arises in the study of the decays $\tau\rightarrow (K, K(1460))\nu_\tau$ and $(K, K(1460))\rightarrow \mu\nu_\mu$ in the extended NJL model. In that setting, the quantities of interest are the weak decay coefficients $F_K$ and $F_K'$, extracted from amplitudes and decay widths involving the kaon, the first radial excitation $K(1460)$, and the axial-vector mesons $K_1(1270)$, $K_1(1400)$, and $K_1(1650)$ [1906.06680].

A common source of ambiguity is that the same letter $K$ indexes conceptually different objects. In the algorithmic setting, it appears in the name of the problem parameter and in the phrase “klam value.” In the hadronic setting, it labels kaonic states and their weak decay coefficients. This suggests that “K-value” is not a single standardized technical term across these literatures.

## 2. Klam value in parameterized complexity

For a parameterized algorithm with running time $O^*(f(k))$, the klam value is defined as the largest integer $k_0$ such that
$$
f(k)<10^{20}\quad\text{for all }0\le k\le k_0.
$$
The $O^*(\cdot)$ notation suppresses factors polynomial in $|V|$ [1502.07725].

The threshold $10^{20}$ was suggested in the very first parameterized-algorithm papers as a yardstick of practical feasibility: a running time $10^{20}$ is already astronomically large, so showing $f(k)<10^{20}$ up to $k=k_0$ attests that the method is “reasonable” for all $k\le k_0$ [1502.07725].

This definition separates the parameterized exponent base from the problem parameter itself. The parameter is $k$, whereas the klam value is a derived benchmark for a specific algorithmic dependence on $k$. In the case of $k$-LST, the 2015 result gives an $O^*(3.188^k)$-time algorithm, which yields a klam value of $39$ because
$$
3.188^{39}<10^{20}<3.188^{40}.
$$

## 3. Historical progression for the $k$-Leaf Spanning Tree problem

The $k$-LST problem has been extensively studied over the past three decades. In 2000, Fellows et al. explicitly asked whether one can obtain a klam value of $50$. From 2000 to 2010, improved branching and measure-and-conquer methods steadily raised the klam value from $17$ to $37$, and the 2010 algorithm of Binkele-Raible and Fernau first reached a klam value of $37$, with running time $O^*(3.46^k)$ [1502.07725].

| Reference | klam value | Running time |
|---|---:|---:|
| Fellows et al. [FL88] | 0 | FPT (non-constructive) |
| Bodlaender [B89] | 1 | $O^*((17k^4)!)$ |
| Fellows et al. [DF95] | 5 | $O^*((2k)^{4k})$ |
| Fellows et al. [FMRS00] | 17 | $O^*(14.23^k)$ |
| Bonsma, Björklund et al. [BBW03] | 20 | $O^*(9.49^k)$ |
| Estivill-Castro et al. [ECFLR05] | 22 | $O^*(8.12^k)$ |
| Bonsma et al. [BBW08] | 24 | $O^*(6.75^k)$ |
| Kneis et al. [KLR08] | 33 | $O^*(4^k)$ |
| Daligault et al. [DGKY08] | 35 | $O^*(3.72^k)$ |
| Binkele-Raible & Fernau [RF10] | 37 | $O^*(3.46^k)$ |
| This paper | 39 | $O^*(3.188^k)$ |

The 2015 result pushes the benchmark from $37$ to $39$. The paper describes this as a solid step toward the open question of reaching klam $50$ [1502.07725].

## 4. The $O^*(3.188^k)$ algorithm and the attainment of klam value 39

The algorithm constructs a partial spanning tree $T\subseteq G$ and maintains three distinguished sets of leaves: $L$ for fixed leaves, $F$ for floating leaves, and $M$ for marked leaves. At each recursive call the instance is represented as $(T,L,M,F)$ together with the remaining quota $k$. The analysis uses the measure
$$
\mu
=
2k
+\frac14|M|
-\Bigl(|L|+|F|+\sum_{i\ge2}(i-1)\,|\text{internal-}T\text{-nodes of degree }i|\Bigr).
$$
Initially $\mu=2k+\tfrac14$, and once $\mu\le0$, the current partial tree already guarantees at least $k$ leaves, so the algorithm accepts in polynomial time [1502.07725].

Each reduction or branching rule either rejects or accepts immediately, or decreases the measure by a positive amount. A branching rule is analyzed through a branching vector $(b_1,b_2,\dots,b_\ell)$, where branch $i$ decreases the measure by $b_i$. If $\alpha$ is the unique positive root of
$$
x^{\max b_i}=\sum_{i=1}^\ell x^{\max b_i-b_i},
$$
then standard recurrence arguments bound the contribution of that rule by $c^\mu$ whenever $\alpha\le c$. Since $\mu\approx 2k$, proving $\alpha\le 3.188^{0.5}$ for all branching rules yields $O^*(3.188^k)$ [1502.07725].

The main new ingredient is the **dependency claim**. Informally, as the algorithm assigns vertices to be leaves or internal vertices, some assignments become impossible because they would contradict earlier branching decisions. The invariant records that when a node $v\in T$ first became internal, its parent $p$ must be internal in any global solution consistent with that branch. If $v$ has a single sibling $s$ in $T$, then $s$ is maintained to be unmarked and to have exactly one attachment path available. This history-dependent invariant is used to prune subbranches that would otherwise be explored [1502.07725].

Two representative rules illustrate the method. In **Rule 8 (Reachability reduction)**, if a vertex $u\notin T$ can be reached only through a single leaf-candidate $v\in M\cup N$, then $v$ must become internal; the algorithm grafts the remaining neighbors of $v$ onto $v$, removes $v$ from $M$, and decreases the measure by at least $\tfrac14$. In **Rule 37 (Two-leaf branching)**, if sibling leaves $v,s\in N$ each have exactly two neighbors outside $T$ with disjoint neighborhoods, the algorithm branches into four exhaustive cases: both leaves, one leaf and one internal in two symmetric ways, and both internal. The branching vector is at least $(5,2,2,2)$, with positive root approximately $1.785$, and $1.785<3.188^{0.5}$ [1502.07725].

By systematically verifying all rules, the worst-case branching factor is bounded by $3.188^{0.5}$. The overall runtime is therefore
$$
O^*\bigl((3.188^{0.5})^{2k+\tfrac14}\bigr)=O^*(3.188^k),
$$
and the algorithm uses polynomial space. Since $3.188^{39}<10^{20}<3.188^{40}$, the algorithm’s klam value is exactly $39$, and the problem admits a klam value of $39$ [1502.07725].

## 5. $F_K$ and $F_K'$ as kaon weak decay coefficients

In the extended NJL model, the quantities $F_K$ and $F_K'$ are the weak decay coefficients of the kaon ground state and the first radial excitation $K(1460)$. The calculation uses constituent quark masses
$$
m_u=m_d=280\,\text{MeV},\qquad m_s=420\,\text{MeV},
$$
together with two cutoffs,
$$
\Lambda_4=1.25\,\text{GeV},\qquad \Lambda_3=1.03\,\text{GeV}.
$$
Ground-state and radially excited pseudoscalar and axial-vector interactions are introduced through the extended NJL Lagrangian, with form factors $f_{us}(k_\perp^2)=1+d_{us}k_\perp^2$, parameter
$$
d_{us}=-1.762\,\text{GeV}^{-2},
$$
and mixing angles
$$
\theta_K=58.11^\circ,\qquad \theta_{K_1A}=85.97^\circ
$$
[1906.06680].

The basic loop integrals are
$$
I_2(m_u,m_s)= -i\frac{N_C}{(2\pi)^4}\int d^4k\,\frac{\Theta(\Lambda_4^2+k^2)}{(m_u^2-k^2)(m_s^2-k^2)}
$$
and
$$
I_2^{f^n}(m_u,m_s)= -i\frac{N_C}{(2\pi)^4}\int d^4k\,
\frac{f_{us}^n(k_\perp^2)\Theta(\Lambda_3^2-k_\perp^2)}{(m_u^2-k^2)(m_s^2-k^2)}.
$$
The ground-state couplings satisfy
$$
g_K=[4I_2(m_u,m_s)]^{-1/2}/\sqrt{Z_K},\qquad g_{K_1A}=\sqrt6\,g_K.
$$

Because the pseudoscalar $K$ can fluctuate via the axial-vector nonet back into $K$, one has
$$
Z_K^{-1}=1-\frac32(m_s+m_u)^2\left[\frac{\sin^2\beta}{M_{K_1(1270)}^2}+\frac{\cos^2\beta}{M_{K_1(1400)}^2}\right],
$$
where the SU(3) mixing angle is $\beta\simeq57^\circ$, with physical masses $M_{1270}=1272\,\text{MeV}$ and $M_{1400}=1403\,\text{MeV}$. Numerically,
$$
Z_K\simeq1.76.
$$

The weak interaction is introduced through
$$
L_{\text{weak}}=\frac{G_F}{\sqrt2}V_{us}\,[\bar\tau\gamma^\mu(1-\gamma_5)\nu_\tau]\,[\bar s\gamma_\mu(1-\gamma_5)u].
$$
Writing the hadronic current as
$$
H_\mu=\sqrt2\,F_K\,p_\mu,
$$
the standard NJL model gives the well-known relation
$$
F_K=\frac{m_s+m_u}{2g_K}.
$$
In the extended variant,
$$
\sqrt2\,F_K=A_C+A_{K_1A}+e^{i\phi}A_{K_1A}'.
$$
Here $A_C$ is the contact term, $A_{K_1A}$ is the contribution from $K_1(1270)$ and $K_1(1400)$, and $A_{K_1A}'$ is the contribution from the radially excited pole $K_1(1650)$ [1906.06680].

## 6. Numerical values, comparison, and significance

For a pseudoscalar meson $P$, the decay width is
$$
\Gamma(\tau\rightarrow P\nu_\tau)=\frac{G_F^2|V_{us}|^2}{16\pi}m_\tau^3F_P^2\left(1-\frac{M_P^2}{m_\tau^2}\right)^2.
$$
Accordingly, once $F_K$ or $F_{K'}$ is known, the width follows immediately, and conversely $F_K$ can be read off from the measured $\tau\rightarrow K\nu$ width [1906.06680].

By varying the phase $\phi$ of the $K_1(1650)$ contribution, the best overall fit to
$$
\Gamma(\tau\rightarrow K\nu)= (1.58\pm0.02)\times10^{-11}\,\text{MeV},
$$
and
$$
\Gamma(K\rightarrow\mu\nu)= (3.379\pm0.006)\times10^{-14}\,\text{MeV}
$$
is obtained for $\phi\simeq102^\circ$. The resulting predictions are:

| Quantity | Value |
|---|---:|
| $\Gamma(\tau\rightarrow K\nu)$ | $1.598\times10^{-11}\,\text{MeV}$ |
| $\Gamma(\tau\rightarrow K'\nu)$ | $1.426\times10^{-13}\,\text{MeV}$ |
| $\Gamma(K\rightarrow\mu\nu)$ | $3.379\times10^{-14}\,\text{MeV}$ |
| $\Gamma(K'\rightarrow\mu\nu)$ | $7.811\times10^{-15}\,\text{MeV}$ |
| $F_K$ | $110.14\,\text{MeV}$ |
| $F_{K'}$ | $29.54\,\text{MeV}$ |

These values reproduce the PDG value
$$
F_K^{\mathrm{exp}}=110.02\,\text{MeV}
$$
to better than $0.2\%$. Lattice QCD with $N_f=2+1+1$ gives $F_K=(109.17$–$110.25)\,\text{MeV}$; light-front quark models give $F_K\approx110$–$114\,\text{MeV}$; and improved holographic models predict $F_K\approx108$–$111\,\text{MeV}$. The NJL result lies in the same $\pm1\%$ window [1906.06680].

The estimated theoretical uncertainty is $\Delta F_K/F_K\simeq5\%$. The sources identified are cutoff dependence, constituent masses, mixing angles, and the relative phase $\phi$ of the radially excited pole. Varying $\Lambda_4$ and $\Lambda_3$ by $\pm5\%$ shifts $F_K$ by at most $\pm3\%$; a $\pm5\%$ change in $m_s$ modifies $F_K$ by $\pm2$–$3\%$; uncertainties of $\pm3^\circ$ in $\beta$, $\theta_K$, and $\theta_{K_1A}$ produce a similar $\pm(1$–$2)\%$ variation; and different choices of $\phi$ can shift $F_{K'}$ by as much as $20\%$, while affecting $F_K$ only at the $<1\%$ level [1906.06680].

The phenomenological implications are stated explicitly. The agreement of $F_K$ with lattice and experiment confirms that extended NJL dynamics correctly encodes chiral-symmetry breaking and the role of axial-vector resonances. Since $F_K$ enters Standard Model predictions for leptonic kaon decays such as $R_K\equiv\Gamma(K\rightarrow e\nu)/\Gamma(K\rightarrow \mu\nu)$, it can serve as an independent theoretical input in tests of lepton flavor universality. The ratio
$$
F_K/F_\pi=1.166
$$
with $F_\pi=94.6\,\text{MeV}$ in the same model agrees at the few-percent level with lattice and ChPT values $\simeq1.195$, and the predicted constant $F_{K'}=29.5\,\text{MeV}$ together with
$$
\Gamma(\tau\rightarrow K(1460)\nu)\approx1.4\times10^{-13}\,\text{MeV}
$$
is proposed as a target for future $\tau$-factory measurements [1906.06680].

Source: https://www.emergentmind.com/topics/k-value