The considerable increase in the data traffic and the number of connected devices, jointly with the development of critical services with high data rate and ultra-reliable low latency, require flexibility, scalability and availability. The network slicing (NS) concept is crucial in managing these diverse traffic demands, dividing the physical network into several logical networks with specific functionality and performance requirements. This work aims to find the most efficient combination of the access network and network slices in 5G heterogeneous scenarios to maximize the quality of service (QoS). We propose the Dynamic Access control and Slice Allocation (DASA) algorithm, focused on users' current needs, their priorities, and network conditions. DASA is based on multi-attribute decision making (MADM) and analytic hierarchy process (AHP) to face the complex problem of network selection. The score function parameters for each candidate network are configured dynamically, taking into account the user preferences and the most sensitive metrics for each service request, such as throughput, delay, jitter, packet loss ratio, and energy consumption. This work presents a new and integral solution that combines software-defined network (SDN) and network function virtualization (NFV) paradigms to improve network performance and user satisfaction. DASA algorithm is validated via simulations and numerical results.