-
Game-Theoretic Analysis of Adversarial Decision Making in a Complex Sociophysical System
Authors:
Andrew C. Cullen,
Tansu Alpcan,
Alexander C. Kalloniatis
Abstract:
We apply Game Theory to a mathematical representation of two competing teams of agents connected within a complex network, where the ability of each side to manoeuvre their resource and degrade that of the other depends on their ability to internally synchronise decision-making while out-pacing the other. Such a representation of an adversarial socio-physical system has application in a range of b…
▽ More
We apply Game Theory to a mathematical representation of two competing teams of agents connected within a complex network, where the ability of each side to manoeuvre their resource and degrade that of the other depends on their ability to internally synchronise decision-making while out-pacing the other. Such a representation of an adversarial socio-physical system has application in a range of business, sporting, and military contexts. Specifically, we unite here two physics-based models, that of Kuramoto to represent decision-making cycles, and an adaptation of a multi-species Lotka-Volterra system for the resource competition. For complex networks we employ variations of the Barabási-Alberts scale-free graph, varying how resources are initially distributed between graph hub and periphery. We adapt as equilibrium solution Nash Dominant Game Pruning as a means of efficiently exploring the dynamical decision tree. Across various scenarios we find Nash solutions where the side initially concentrating resources in the periphery can sustain competition to achieve victory except when asymmetries exist between the two. When structural advantage is limited we find that agility in how the victor stays ahead of decision-state of the other becomes critical.
△ Less
Submitted 4 December, 2023; v1 submitted 27 November, 2023;
originally announced November 2023.
-
Unifying warfighting functions in mathematical modelling: combat, manoeuvre, and C2
Authors:
Ryan Ahern,
Mathew Zuparic,
Keeley Hoek,
Alexander Kalloniatis
Abstract:
The outcomes of warfare have rarely only been characterised by the quantity and quality of individual combatant force elements. The ability to manoeuvre and adapt across force elements through effective Command and Control (C2) can allow smaller or weaker forces to overcome an adversary with greater resource and fire-power. In this paper, we combine the classic Lanchester combat model with the Kur…
▽ More
The outcomes of warfare have rarely only been characterised by the quantity and quality of individual combatant force elements. The ability to manoeuvre and adapt across force elements through effective Command and Control (C2) can allow smaller or weaker forces to overcome an adversary with greater resource and fire-power. In this paper, we combine the classic Lanchester combat model with the Kuramoto-Sakaguchi model for phase oscillators on a network to create a flexible Networked-Lanchester-C2 representation of force-on-force military engagement. The mathematical model thus unifies three of the military warfighting `functions': fires, manoeuvre and C2. We consider three illustrative use-cases, and show that an analytical treatment of a reduced model characterises global effects in the full system. For inhomogeneous forces we observe that with appropriate balance between internal organisational coupling, resource manoeuvrability and even weaker lethality the force can be adaptive to overcome an initially stronger adversary.
△ Less
Submitted 24 January, 2023;
originally announced January 2023.
-
Adversarial Decisions on Complex Dynamical Systems using Game Theory
Authors:
Andrew C. Cullen,
Tansu Alpcan,
Alexander C. Kalloniatis
Abstract:
We apply computational Game Theory to a unification of physics-based models that represent decision-making across a number of agents within both cooperative and competitive processes. Here the competitors try to both positively influence their own returns, while negatively affecting those of their competitors. Modelling these interactions with the so-called Boyd-Kuramoto-Lanchester (BKL) complex d…
▽ More
We apply computational Game Theory to a unification of physics-based models that represent decision-making across a number of agents within both cooperative and competitive processes. Here the competitors try to both positively influence their own returns, while negatively affecting those of their competitors. Modelling these interactions with the so-called Boyd-Kuramoto-Lanchester (BKL) complex dynamical system model yields results that can be applied to business, gaming and security contexts. This paper studies a class of decision problems on the BKL model, where a large set of coupled, switching dynamical systems are analysed using game-theoretic methods.
Due to their size, the computational cost of solving these BKL games becomes the dominant factor in the solution process. To resolve this, we introduce a novel Nash Dominant solver, which is both numerically efficient and exact. The performance of this new solution technique is compared to traditional exact solvers, which traverse the entire game tree, as well as to approximate solvers such as Myopic and Monte Carlo Tree Search (MCTS). These techniques are assessed, and used to gain insights into both nonlinear dynamical systems and strategic decision making in adversarial environments.
△ Less
Submitted 28 January, 2022;
originally announced January 2022.
-
Two network Kuramoto-Sakaguchi model under tempered stable Lévy noise
Authors:
Alexander Kalloniatis,
Timothy McLennan-Smith,
Dale Roberts,
Mathew Zuparic
Abstract:
We examine a model of two interacting populations of phase oscillators labelled `Blue' and `Red'. To this we apply tempered stable Lévy noise, a generalisation of Gaussian noise where the heaviness of the tails parametrised by a power law exponent $α$ can be controlled by a tempering parameter $λ$. This system models competitive dynamics, where each population seeks both internal phase synchronisa…
▽ More
We examine a model of two interacting populations of phase oscillators labelled `Blue' and `Red'. To this we apply tempered stable Lévy noise, a generalisation of Gaussian noise where the heaviness of the tails parametrised by a power law exponent $α$ can be controlled by a tempering parameter $λ$. This system models competitive dynamics, where each population seeks both internal phase synchronisation and a phase advantage with respect to the other population, subject to exogenous stochastic shocks. We study the system from an analytic and numerical point of view to understand how the phase lag values and the shape of the noise distribution can lead to steady or noisy behaviour. Comparing the analytic and numerical studies shows that the bulk behaviour of the system can be effectively described by dynamics in the presence of tilted ratchet potentials. Generally, changes in $α$ away from the Gaussian noise limit, $1< α< 2$, disrupts the locking between Blue and Red, while increasing $λ$ acts to restore it. However we observe that with further decreases of $α$ to small values, $α\ll 1$, with $λ\neq 0$, locking between Blue and Red may be restored. This is seen analytically in a restoration of metastability through the ratchet mechanism, and numerically in transitions between periodic and noisy regions in a fitness landscape using a measure of noise. This non-monotonic transition back to an ordered regime is surprising for a linear variation of a parameter such as the power law exponent and provides a novel mechanism for guiding the collective behaviour of such a complex competitive dynamical system.
△ Less
Submitted 28 April, 2021;
originally announced April 2021.
-
Adversarial decision strategies in multiple network phased oscillators: the Blue-Green-Red Kuramoto-Sakaguchi model
Authors:
Mathew Zuparic,
Maia Angelova,
Ye Zhu,
Alexander Kalloniatis
Abstract:
We consider a model of three interacting sets of decision-making agents, labeled Blue, Green and Red, represented as coupled phased oscillators subject to frustrated synchronisation dynamics. The agents are coupled on three networks of differing topologies, with interactions modulated by different cross-population frustrations, internal and cross-network couplings. The intent of the dynamic model…
▽ More
We consider a model of three interacting sets of decision-making agents, labeled Blue, Green and Red, represented as coupled phased oscillators subject to frustrated synchronisation dynamics. The agents are coupled on three networks of differing topologies, with interactions modulated by different cross-population frustrations, internal and cross-network couplings. The intent of the dynamic model is to examine the degree to which two of the groups of decision-makers, Blue and Red, are able to realise a strategy of being ahead of each others' decision-making cycle while internally seeking synchronisation of this process -- all in the context of further interactions with the third population, Green. To enable this analysis, we perform a significant dimensional reduction approximation and stability analysis. We compare this to a numerical solution for a range of internal and cross-network coupling parameters to investigate various synchronisation regimes and critical thresholds. The comparison reveals good agreement for appropriate parameter ranges. Performing parameter sweeps, we reveal that Blue's pursuit of a strategy of staying too-far ahead of Red's decision cycles triggers a second-order effect of the Green population being ahead of Blue's cycles. This behaviour has implications for the dynamics of multiple interacting social groups with both cooperative and competitive processes.
△ Less
Submitted 19 November, 2020;
originally announced November 2020.
-
Gaussian noise and the two-network frustrated Kuramoto model
Authors:
A. B. Holder,
M. L. Zuparic,
A. C. Kalloniatis
Abstract:
We examine analytically and numerically a variant of the stochastic Kuramoto model for phase oscillators coupled on a general network. Two populations of phased oscillators are considered, labelled `Blue' and `Red', each with their respective networks, internal and external couplings, natural frequencies, and frustration parameters in the dynamical interactions of the phases. We disentagle the dif…
▽ More
We examine analytically and numerically a variant of the stochastic Kuramoto model for phase oscillators coupled on a general network. Two populations of phased oscillators are considered, labelled `Blue' and `Red', each with their respective networks, internal and external couplings, natural frequencies, and frustration parameters in the dynamical interactions of the phases. We disentagle the different ways that additive Gaussian noise may influence the dynamics by applying it separately on zero modes or normal modes corresponding to a Laplacian decomposition for the sub-graphs for Blue and Red. Under the linearisation ansatz that the oscillators of each respective network remain relatively phase-sychronised centroids or clusters, we are able to obtain simple closed-form expressions using the Fokker-Planck approach for the dynamics of the average angle of the two centroids. In some cases, this leads to subtle effects of metastability that we may analytically describe using the theory of ratchet potentials. These considerations are extended to a regime where one of the populations has fragmented in two. The analytic expressions we derive largely predict the dynamics of the non-linear system seen in numerical simulation. In particular, we find that noise acting on a more tightly coupled population allows for improved synchronisation of the other population where deterministically it is fragmented.
△ Less
Submitted 29 September, 2016;
originally announced September 2016.
-
Fixed points and stability in the two-network frustrated Kuramoto model
Authors:
A. C. Kalloniatis,
M. L. Zuparic
Abstract:
We examine a modification of the Kuramoto model for phase oscillators coupled on a network. Here, two populations of oscillators are considered, each with different network topologies, internal and cross-network couplings and frequencies. Additionally, frustration parameters for the interactions of the cross-network phases are introduced. This may be regarded as a model of competing populations: i…
▽ More
We examine a modification of the Kuramoto model for phase oscillators coupled on a network. Here, two populations of oscillators are considered, each with different network topologies, internal and cross-network couplings and frequencies. Additionally, frustration parameters for the interactions of the cross-network phases are introduced. This may be regarded as a model of competing populations: internal to any one network phase synchronisation is a target state, while externally one or both populations seek to frequency synchronise to a phase in relation to the competitor. We conduct fixed point analyses for two regimes: one, where internal phase synchronisation occurs for each population with the potential for instability in the phase of one population in relation to the other; the second where one part of a population remains fixed in phase in relation to the other population, but where instability may occur within the first population leading to `fragmentation'. We compare analytic results to numerical solutions for the system at various critical thresholds.
△ Less
Submitted 16 November, 2015;
originally announced November 2015.
-
Stochastic (in)stability of synchronisation of oscillators on networks
Authors:
Mathew Zuparic,
Alexander C. Kalloniatis
Abstract:
We consider the influence of correlated noise on the stability of synchronisation of oscillators on a general network using the Kuramoto model for coupled phases $θ_i$. Near the fixed point $θ_i \approx θ_j \ \forall i,j$ the impact of the noise is analysed through the Fokker-Planck equation. We deem the stochastic system to be `weakly unstable' if the Mean First Passage Time for the system to dri…
▽ More
We consider the influence of correlated noise on the stability of synchronisation of oscillators on a general network using the Kuramoto model for coupled phases $θ_i$. Near the fixed point $θ_i \approx θ_j \ \forall i,j$ the impact of the noise is analysed through the Fokker-Planck equation. We deem the stochastic system to be `weakly unstable' if the Mean First Passage Time for the system to drift outside the fixed point basin of attraction is less than the time for which the noise is sustained. We argue that a Mean First Passage Time, computed near the phase synchronised fixed point, gives a useful lower bound on the tolerance of the system to noise. Applying the saddle point approximation, we analytically derive general thresholds for the noise parameters for weak stochastic stability. We illustrate this by numerically solving the full Kuramoto model in the presence of noise for an example complex network.
△ Less
Submitted 25 April, 2013; v1 submitted 4 September, 2011;
originally announced September 2011.
-
From incoherence to synchronicity in the network Kuramoto model
Authors:
Alexander C. Kalloniatis
Abstract:
We study the synchronisation properties of the Kuramoto model of coupled phase oscillators on a general network. Here we distinguish the ability of such a system to self-synchronise from the stability of this behaviour. While self-synchronisation is a consequence of genuine non-perturbative dynamics, the stability in dynamical systems is usually accessible by fluctuations about a fixed point, he…
▽ More
We study the synchronisation properties of the Kuramoto model of coupled phase oscillators on a general network. Here we distinguish the ability of such a system to self-synchronise from the stability of this behaviour. While self-synchronisation is a consequence of genuine non-perturbative dynamics, the stability in dynamical systems is usually accessible by fluctuations about a fixed point, here taken to be the synchronised solution. We examine this problem in terms of modes of the graph Laplacian, by which the absolute Lyapunov stability of the synchronised fixed point is readily demonstrated. Departures from stability are seen to arise at the next order in fluctuations where the dynamical equations resemble those for species population models, the logistic and Lotka-Volterra equations. Methods from these systems are exploited to analytically derive new critical couplings signalling deviation from classical stability. We observe in some cases an intermediate regime of behaviour, between incoherence and synchronisation, where system wide periodic behaviours are exhibited. We discuss these results in light of simulations.
△ Less
Submitted 1 September, 2009;
originally announced September 2009.
