On the ambiguity of determination of interfering resonances parameters

The general form of solutions for parameters of interfering Breit-Wigner resonances is found. The number of solutions is determined by the properties of roots of corresponding characteristic equation and does not exceed $2^{N-1}$, where $N$ is the number of resonances. For resonances of more complicated form, provided that their amplitudes satisfy certain conditions, for any $N\ge2$ multiple solutions also exist.

💡 Research Summary

The paper addresses a subtle but crucial problem that arises when several resonant states interfere in scattering or decay processes: the extraction of the underlying resonance parameters (mass, width, coupling strength, phase) from measured cross‑section data is not unique. Starting from the standard Breit‑Wigner (BW) description, the authors consider a superposition of N BW amplitudes, each characterized by a complex coupling (c_i e^{i\phi_i}) and a pole at (M_i - i\Gamma_i/2). The observable quantity is the modulus squared of the total amplitude, (\sigma(E)=|A(E)|^{2}), which is a highly non‑linear function of the parameters.

By algebraically clearing denominators, the authors rewrite the squared amplitude as a ratio of two real‑valued polynomials in the energy variable (E). The numerator and denominator are constructed from the product of the amplitude with its complex conjugate, leading to a characteristic polynomial equation (F(z)=0) whose roots (z_k) encode the possible parameter sets that reproduce the same (\sigma(E)). The degree of this polynomial is (2N-2); consequently, there can be at most (2N-2) distinct roots. Each root gives rise to two possible choices of the complex couplings (essentially a sign ambiguity), so the total number of distinct solutions is bounded by (2^{k-1}), where (k) is the number of relevant roots (with (k\le N)). In the generic situation where all N resonances contribute independently, the maximal number of physically distinct parameter sets is therefore (2^{N-1}).

The analysis is then extended beyond the simple BW form. The authors consider more general resonant amplitudes that may have energy‑dependent widths, background terms, or coupled‑channel effects, provided the total amplitude can still be expressed as a linear combination of known complex functions (f_i(E)) with constant coefficients. Under this mild condition the same polynomial construction applies, and the same bound on the number of solutions holds. Hence, for any system with (N\ge2) interfering resonances that satisfy the stated linearity condition, multiple mathematically equivalent parameter sets inevitably exist.

From a practical standpoint, the paper warns that a standard chi‑square minimisation that starts from a single initial guess may converge to any one of these equivalent minima, potentially leading to ambiguous physical interpretations. To mitigate this, the authors recommend: (1) incorporating external constraints such as phase information from other channels, theoretical sum rules, or known symmetry relations; (2) employing global optimisation techniques (genetic algorithms, simulated annealing, Markov‑Chain Monte‑Carlo) with a broad set of initial conditions to map all minima; (3) using additional observables (angular distributions, polarization data) that are sensitive to the relative phases and can discriminate among the solutions.

In summary, the work provides a rigorous mathematical framework that quantifies the intrinsic ambiguity in extracting interfering resonance parameters. It demonstrates that the ambiguity is not an artifact of experimental noise or fitting algorithms but a direct consequence of the algebraic structure of the problem. The results have immediate implications for high‑precision analyses in particle and nuclear physics, where overlapping resonances are common, and they underscore the necessity of supplementing fits with independent physical constraints to obtain a unique and reliable set of resonance parameters.