[en] Tungsten (W) is the main candidate for the first wall of a reactor due to its robustness against physical sputtering by the plasma ions, however, when W reaches concentrations of 10^-4 in the plasma, it causes unduly large plasma cooling by radiation. This implies restrictive impurity control for W, which needs reliable diagnostic by plasma spectroscopy. The most intense spectral lines of highly ionized W are emitted in the VUV and soft X-ray range. To perform calculations on atomic data the code packages incorporated in the ADAS project are used. The electronic structure of nearly all W-ions is calculated by the Cowan-code (Hartree-Fock algorithm). In a second step, the cross sections for electron impact excitation are evaluated via the Cowan-code using the plane wave Born-approximation. A detailed collisional-radiative model is employed to calculate the model-spectra for each ion in equilibrium. Finally, ionization and recombination rates of W are evaluated by semi-empirical formulae, which make use of the electronic structure calculations of the Cowan-code. All atomic data are confronted with experimental measurements from the Garching tokamak ASDEX Upgrade and the Berlin electron-beam ion trap (EBIT). The experimental investigations extend up to 5 keV electron temperatures, which is the maximum of the routine operation at ASDEX Upgrade. 'Impurity accumulation', which is characterized by a strong peaking of the impurity density profile, enables unique investigations on the fractional abundance of Ag-like W²⁷⁺ up to Co-like W⁴⁷⁺. The recombination rates for few states are corrected empirically satisfying boundary conditions which arise from experimental evidence. Both spectral features have been studied also for isoelectronic sequences by injecting the impurities hafnium, tantalum, rhenium, gold, lead and bismuth. Additionally, xenon is targeted by the same code packages, as xenon might be injected in future experiments or a reactor for intentional plasma cooling. Predictions on radiative plasma cooling (cooling factor CF) have been based up to now on the rough 'Average Ion Model' (AIM) and a further result of the work is the analysis of plasma cooling with the outlined, superior model. All data, which are benchmarked by experimental spectra, are used to calculate the CF of the high-Z elements. (Orig.)