The paper looks into 5G network slicing in comprehensive detail, addressing challenges from resource allocation, scheduling, security, and even specific applications. In order to enhance the performance and distribution of resources across numerous vertical businesses, it presents network slicing principles like Use-case Specific and Sub Network Slicing. Beamforming is used in a channel-aware and SLA-aware scheduling system that optimizes resource usage and reduces resource block consumption by 60.9%. Smart grid applications are examined, including techniques such as NOMA for power minimization in eMBB and URLLC traffic, various 5G radio access numerologies, and multi-dimensional Markov models to evaluate transient performance. While multi-agent reinforcement learning techniques like COQRA significantly increase latency, throughput, and packet drop rates, knowledge transfer-based resource allocation (KTRA) performs better in terms of latency and throughput than Q-learning. The T-S3RAmodel that has been confirmed on a network simulator with 250 nodes, addresses security by employing Renyi entropy to predict DDoS attacks and Boolean logic to authenticate devices. Through NSBchain, the study also makes use of blockchain technology for innovative business models and safe resource allocation. While DMUSO optimises scheduling, heuristic algorithms like MCRA and FLDRA enhance resource use. The Network Slicing models taxonomy focuses on how SDN and NFV support overcoming architectural defects in 5G to provide customized services at fast data rates with minimal latency. Along with offering innovative solutions and architectural frameworks for the various 5G use cases, the study also presents future research prospects in resource allocation and service quality enhancements.