Competitive Three-Species Lotka-Volterra Dynamics

• The competitive three-species Lotka–Volterra model is a mathematical framework defined by ODEs that capture interspecific and intraspecific competition via growth rates and a competition matrix.
• It enables explicit classifications of equilibria, permanence conditions, and rich dynamical regimes including bistability, limit cycles, and heteroclinic cycles.
• Extensions incorporating seasonal forcing and spatial diffusion provide rigorous bifurcation analyses and insights into community assembly and coexistence.

The competitive three-species Lotka–Volterra model is a canonical mathematical framework for exploring the population dynamics of three species in pairwise competition. This model has played a central role in ecology, mathematical biology, and nonlinear dynamics since its inception, with modern theoretical work delivering comprehensive classifications, global stability results, explicit Lyapunov constructions, and bifurcation analyses for both classical and seasonally forced variations. Its phase portrait admits rich behavior, from unique globally attracting equilibria to bistability, heteroclinic cycles, stable limit cycles, and quasi-periodic orbits, with precise dynamics determined by competition coefficients, growth parameters, and the symmetry structure of interspecific interactions.

1. Model Definition and Governing Equations

The general competitive three-species Lotka–Volterra system is formulated as a system of first-order ODEs for population densities $x_i(t)\ge 0$: $$\dot x_i = x_i\left( r_i - \sum_{j=1}^3 b_{ij} x_j \right), \quad i = 1,2,3,$$ where $r_i > 0$ are intrinsic growth rates and $b_{ij}\geq 0$ measure the per-capita competitive effect of species $j$ on species $i$ (Champagnat et al., 2010). The competition matrix $B = (b_{ij})$ encapsulates both intraspecific ($i = j$) and interspecific ($i \ne j$) competitive interactions. Extensions include piecewise (seasonally) forced systems with alternating “bad” (decay) and “good” (Lotka–Volterra) seasons (Niu et al., 2023, Niu et al., 29 Feb 2024), spatially structured reaction–diffusion models, and variants incorporating cyclic dominance or interface dynamics (Girardin, 2018, Álvarez-Caudevilla et al., 6 Aug 2024).

2. Equilibria, Fixed Points, and Permanence

Steady states correspond to solutions of the linear system $B x^* = r$, with $x^* = B^{-1} r$ when $B$ is invertible and $x^* > 0$ (Champagnat et al., 2010). Under detailed-balance or symmetrized competition (i.e., existence of $C_i>0$ so that $C_i b_{ij} = C_j b_{ji}$ for all $i, j$) and positive definiteness of $M = (C_i b_{ij})$ on $\mathbb{R}^3 \setminus \{0\}$, a unique strictly positive equilibrium $x^*$ exists and is globally attracting. In the context of seasonally forced systems, the classification of extinction, coexistence (permanence), or boundary attractors is achieved via a finite combinatorial catalogue of boundary classes, with explicit parameter inequalities dictating permanence or impermanence (Niu et al., 2023, Niu et al., 29 Feb 2024). A carrying simplex, an invariant codimension-one hypersurface in $\mathbb{R}^3_+$, organizes global attractor structures for discrete-time Poincaré maps (Niu et al., 2023, Niu et al., 29 Feb 2024).

Permanence—uniform persistence of all three species for all time—holds in precisely those parameter classes where explicit weighted sums of boundary Floquet multipliers are positive at all possible boundary fixed points (Niu et al., 29 Feb 2024). This framework subsumes classical coexistence, with permanence classes coinciding with interior fixed point existence and additional algebraic conditions.

3. Global Dynamics and Lyapunov Structure

The global phase portrait is governed by the competition matrix $B$, its symmetry properties, and the existence of Lyapunov functions. Under detailed-balance (weighted-symmetry) and $M$ positive definite, the system admits a strict Lyapunov function of the form

$$F(x) = \sum_{i=1}^3 C_i (x_i \ln x_i - x_i) + \frac{1}{2} \sum_{i,j=1}^3 C_i b_{ij} x_i x_j$$

with $\dot F(x) \le 0$ and equality iff $x = x^*$. This construction yields global convergence to the unique equilibrium, precluding cycles, heteroclinic networks, or more complicated attractors (Champagnat et al., 2010).

Relaxing symmetry or introducing certain cyclic competition coefficients enables richer dynamics. The May–Leonard variant admits heteroclinic cycles or neutrally stable periodic orbits at critical parameter values (Tang et al., 2012). Explicit Lyapunov constructions demonstrate the dichotomy:

• Generalized gradient cases: strictly decreasing Lyapunov function, no cycles, unique attractor or bistability, depending on the Hessian signature at $x^*$;
• Non-gradient (cyclic) cases: rotational flows in phase space, supporting heteroclinic cycles or limit cycles at criticality.

For forced (seasonal) systems, the global dynamics reduce to attractors on the carrying simplex, whose topology and content varies across 33 explicit boundary classes; these may yield unique interior attractors, multi-stability, heteroclinic cycles, or attracting invariant closed curves supporting quasi-periodic behavior (Niu et al., 29 Feb 2024, Niu et al., 2023).

4. Oscillatory and Spatial Dynamics

For three-species Lotka–Volterra competition–diffusion systems, additional spatiotemporal complexity arises. Explicit construction of three-component KPP–Lotka–Volterra systems demonstrates the occurrence of stable limit cycles (Hopf bifurcation scenario), periodic wave trains, monotone traveling fronts, and point-to-periodic (wave-train) invasion phenomena (Girardin, 2018). In well-mixed ODEs, a two-dimensional unstable manifold around the coexistence equilibrium supports a globally stable limit cycle post-Hopf, with population dynamics exhibiting cyclic predator–prey structure near the steady state.

On inclusion of space (diffusion), propagation phenomena—including traveling waves, periodic invasion fronts, and pulsating terraces—are explicitly constructed or characterized via rigorous spectral and comparison results (Girardin, 2018). Notably, pattern formation and oscillatory dynamics are typically suppressed with strictly local interactions unless explicit mobility or long-range movement is present (Palombi et al., 2019).

5. Classification, Bifurcation, and Nonuniqueness

A complete global classification of three-species competitive Lotka–Volterra systems with generic parameters (including time-periodic seasonal succession) is obtained via reduction to boundary dynamics on the carrying simplex and enumeration into 33 equivalence classes, uniquely determined by sign patterns of derived inequalities in the parameters (Niu et al., 2023, Niu et al., 29 Feb 2024). Within these, regimes exist of:

• Trivial dynamics: all orbits tend to a boundary equilibrium;
• Global permanence with unique attractor: all strongly positive orbits converge to the interior fixed point;
• Heteroclinic cycles or quasi-periodic solutions: supported in certain classes (notably class 27 and others), with explicit algebraic criteria (e.g., sign of $\vartheta$) dictating the topology of attractors.

Uniqueness of interior fixed points may fail—unlike in the two-species case—leading to multi-attractor dynamics, as numerically demonstrated in specific classes (Niu et al., 29 Feb 2024). Even in permanence regimes, the attracting set can be a closed invariant curve rather than a point, supporting quasi-periodicity.

6. Extensions: Interfaces, Spatial Fragmentation, and Further Models

Recent extensions analyze three-species Lotka–Volterra systems on spatial domains partitioned by membranes/interfaces, where interaction is restricted to region boundaries (Álvarez-Caudevilla et al., 6 Aug 2024). This setting introduces new bifurcation phenomena: breaking of the classical competitive exclusion principle, source–sink dynamics (e.g., persistence of populations in "sink" habitats due to boundary inflow), and coexistence “islands” in parameter space that are strongly shaped by interface permeability and landscape geometry—not encountered in the homogeneous model.

Cyclic variants and models with higher-order (e.g., three-agent) interactions elaborate further on these themes, providing a rich taxonomy of dynamical regimes, explicit Hopf bifurcation points, and spatial–temporal pattern formation—particularly in multi-patch scenarios with explicit mobility (Palombi et al., 2019, Girardin, 2018).

7. Biological and Evolutionary Implications

The rigorous analysis of the competitive three-species Lotka–Volterra model formalizes and quantifies several foundational principles:

• Competitive coexistence: under symmetric/weighted-symmetric competition, robust coexistence—interpreted as Evolutionarily Stable Strategy (ESS)—is guaranteed (Champagnat et al., 2010).
• Exclusion and rescue: spatial structure (interfaces, fragmented habitats) and time-variation (seasonal succession) can both enforce or prevent coexistence, depending in finely tuned ways on parameter regimes (Álvarez-Caudevilla et al., 6 Aug 2024, Niu et al., 2023).
• Community assembly: in adaptive-dynamics and evolutionary models, symmetric Lotka–Volterra interactions provide the canonical approximation of stable resident community states in the limit of large population and rare mutations (Champagnat et al., 2010).

The direct and indirect consequences of system symmetry, boundary structure, and external forcing are now comprehensively classified for three-species competition, forming a mature and robust branch of theoretical population biology.

---

Key References:

• "Convergence to equilibrium in competitive Lotka-Volterra equations" (Champagnat et al., 2010)
• "Global dynamics of three-dimensional Lotka-Volterra competition models with seasonal succession: I. Classification of dynamics" (Niu et al., 2023)
• "Classification of permanence and impermanence for a Lotka-Volterra model of three competing species with seasonal succession" (Niu et al., 29 Feb 2024)
• "Two components is too simple: an example of oscillatory Fisher--KPP system with three components" (Girardin, 2018)
• "Determine dynamical behaviors by the Lyapunov function in competitive Lotka-Volterra systems" (Tang et al., 2012)
• "A three population Lotka-Volterra competition model with two populations interacting through an interface" (Álvarez-Caudevilla et al., 6 Aug 2024)
• "Coevolutionary dynamics of a variant of the cyclic Lotka-Volterra model with three-agent interactions" (Palombi et al., 2019)
• "General Properties of a System of $S$ Species Competing Pairwise" (Zia, 2010)