Late-stage reversal of negative information diffusion driven by memory accumulation under low infection on simplicial complexes

Negative information diffusion on complex networks depends on the interaction between network structure and dynamics. Considering the higher-order topology and memory effects, this paper constructs a fractional-order negative information diffusion model for both homogeneous and heterogeneous networks, through the framework of simplicial complexes. The existence and stability of equilibrium points and thresholds are analytically derived for each network type, with the bifurcation conditions derived for bi-stability emergence. Optimal control strategies are developed for both models through constrained optimization of the system gains. Experiments conducted on synthetic and real simplicial complexes demonstrate that bi-stability is induced by higher-order interactions but destroyed by fractional order. Notably, at elevated infection rates, a pronounced inverse correlation is observed between the memory effect and diffusion scale, whereas under lower transmission conditions, enhanced memory effects induce a delayed reversal of information propagation, thereby elevating the control costs and implementation challenges. Higher-order interactions can inhibit this reversal. Additionally, the dominance of cognitive reinforcement and corrective strategies in synergistic control relies on the network degree distribution. This research presents a promising methodology for analyzing and controlling negative information diffusion.