The network dynamics captures many fundamental processes involving F-actin, including polymerization, depolymerization, de novo filament nucleation, branching from mother filaments, detachment of branches, and dissociation of filaments due to depolymerization. This model allows visualization of detailed actin network structure within actin waves and is more amenable to further analysis than stochastic simulations previously studied by Carlsson. Simulations of this model reveal possible mechanisms that drive propagation of actin waves. By restricting filament orientation to the vertical direction, we demonstrated that propagation of actin waves could be caused by diffusion of cytosolic proteins which regulate filament nucleation, although in reality the propagation speed may depend on both diffusion of the proteins and polymerization of forward-oriented filaments. Our model also highlights the role of PI3K in initiation and propagation of actin waves, as the autocatalytic cooperativity introduced by the positive feedback through PIP3 is crucial for accumulation of local F-actin density. It is suggested that a delicate balance between this positive feedback and a negative feedback that localizes cell protrusions has to be reached for efficient control of cell movement. In the excitable media framework, which is a basis for existing Benzethonium Chloride actin-wave models, decay of the wave back is caused by activity of the slow inhibitor. Although it suggests annihilation of colliding wave fronts, it fails to explain PIP3 activity behind actin waves and does not allow retraction of actin waves. In contrast, our model suggests that decay of F-actin intensity at the back of actin waves may be caused by local scarcity of free cytosolic molecules such as Arp2/3 and G-actin, which are fundamental components of the actin network. The gradual decay of wave backs by depolymerization of exposed filaments leads to the observed threedimensional structure of actin waves. This mechanism explains the experimental observation that PIP3 is localized in the region enclosed by wave fronts, and has the largest gradient at the wave fronts. It also predicts retraction and formation of new wave fronts when PIP3 activity is locally disrupted by PTEN. Moreover, the scarcity-based decay suggests annihilation of wave fronts as they collide. Similarly, it has been reported that local scarcity of Arp2/3 or G-actin leads to retraction behind protrusion waves at the leading edge of epithelial cells. In this work, filament capping, severing, and disintegration is omitted, and, to obtain adequate turnover of barbed ends to cause decay of the wave back, a relatively high depolymerization rate is needed. However, one can view this higher activity as reflective of these other processes and think of them as implicitly integrated in the depolymerization rate used in the simulations. Numerical simulations of a stochastic model that also includes barbed-end capping suggests that it facilitates decay of the back of actin waves. Many molecules including Arp2/3, MyoB, CARMIL, coronin, and, in fibroblasts, integrin are Pancuronium dibromide associated with actin waves. Their localization is either slightly leading, slightly lagging, or coinciding with that of F-actin. Our model suggests that their localization is dictated by their roles in regulation of the actin network. Arp2/3 is an integral component of the network, responsible for branching, and thus colocalizes with F-actin. On the other hand, coronin regulates debranching and appears to decorate filaments at the top, slightly lagging the wave fronts. Although it is unclear what gives rise to the complex dynamics of PTEN in actin waves, it plays a crucial role in the actin wave dynamics.