The cobweb model of competitive market dynamics has been examined in the form of system dynamics model. Separation of the structure elements and introduction of anticipative hyperincursive algorithm was used for transformation of the classical cobweb model to the accelerator based one. The cyclical response of the system that depends on the demand~supply parameters and eigenvalues of the characteristic equation has been numerically examined. The concept of parameter differentiation and time response of the system is transformed to the periodicity concept where periodicity is the main, driven property of the model. As such this is the key attribute in complex discrete agent-based adaptive anticipatory models. The periodic conditions of the model have been analytically determined by the application of z-transform. The periodicity conditions of the initial map have been preserved in the nonlinear case. By the application of the Lyapunov exponents several stability regions of the nonlinear model were numerically determined.