Predictive modelling of plasma profiles is an essential part of ongoing research in tokamak plasmas, required for a successful realization of future fusion reactors. This thesis focuses on upgrading the RAPTOR code to extend the area of its applicability for plasma modelling and scenario development. RAPTOR is a light and fast simulator, solving radial transport equations, developed for plasma real-time control. This thesis also demonstrates new strategy for ramp-down optimization. The RAPTOR transport model has been extended to take into account the influence of the time-varying plasma equilibrium geometry and background kinetic profiles on the evolution of the predicted plasma profiles. It allows to get more realistic predictions of the plasma state in case of rapid changes in the plasma shape and equilibrium. Also transport equations for the ion temperature and plasma particles (electrons and ions) have been implemented in the code. Benchmarks have been performed with more sophisticated transport ASTRA and CRONOS codes and with prescribed data for the particle transport in ITER. With successful benchmarks, we confirm that the new transport equations are solved correctly. A new ad-hoc transport model based on constant gradients for core and pedestal regions, that is suitable for simulations of transition between H (high) and L (low confinement) modes, has been implemented into RAPTOR. This model assumes ``stiffness'' of the plasma profiles in the core region, reflecting their relatively weak reaction to changes in the heat flux. Only few transport model parameters have to be prescribed. They are validated with predictive simulations of the time evolution of plasma profiles for TCV, ASDEX Upgrade and JET plasmas. We demonstrate the capabilities of RAPTOR for fast and realistic predictions of plasma state over the entire plasma discharges, i.e. from ramp-up to ramp-down. We have defined characteristic gradients in the ``stiff'' region for each machine and L/H confinement modes and have obtained a very good agreement with experimental measurements. We have also demonstrated several special cases, where the obtained set of the transport parameters does not work, and proposed possible solutions of the problems. An optimization procedure for the plasma ramp-down phase has been developed during this work. Nondisruptive termination scenarios are necessary for safe operation of ITER, since it can withstand only a limited amount of plasma disruptions. Automatic optimization algorithms can be applied for searching the optimal ramp-down trajectory. With RAPTOR, optimization results are obtained in a reasonable time (hours). We define the goal of the optimization as ramping down the plasma current as fast as possible while avoiding any disruptions caused by reaching physical or technical limits. Physical constraints are relevant for most tokamaks, others are technical and related to the specific tokamaks. We show how different goals and constraints can easily be included or updated in order to simulate a new machine. A proper plasma shaping during the current ramp-down can reduce significantly the plasma internal inductance, improving its vertical stability. Specific heating scenarios allow to reduce the drop in βpol during H-L transition, which is important for plasma MHD stability. Results of numerical and experimental ramp-down studies for TCV, AUG and JET plasmas are presented.