This paper explores the potential of generative neural network simulations, focusing on the application of reinforcement learning methods and neural world models for creating interactive worlds. Key achievements in agent training using reinforcement learning are discussed. Special attention is given to neural world models, as well as generative models such as Oasis, DIAMOND, Genie, and GameNGen, which employ diffusion networks to generate realistic and interactive game worlds. The opportunities and limitations of generative simulation models are examined, including issues related to error accumulation and memory constraints, and their impact on the quality of generation. The conclusion presents suggestions for future research directions.

To disclose the concept of game balance and to develop an approach to automate the routine when working with game economics, methods of data analysis and experimentation were applied. According to the results of the analysis of existing definitions, a special case of mathematical balance and a generalized design definition of game balance were singled out. By parsing the existing approaches for balancing and testing games, a vision of our own solution was developed. Based on the findings, an approach for automating balance within a game prototype generator has been proposed. As an intermediate result, an updated structure and operation procedure of the game prototype generation tool were presented. The prospects for further development in this direction are given.