This document discusses a proposed spiking neural network (SNN) approach for dynamic resource allocation in software defined networking (SDN). The SNN would allocate bandwidth resources among network slices based on demand. It analyzes past traffic data using an online training method to continuously learn and adapt allocations in real-time. The SNN architecture has input, hidden and output layers to estimate traffic. Simulation results showed a 98.8% benefit to network throughput from this approach. Machine learning and SNNs are well-suited for dynamic resource optimization in 5G networks due to their ability to analyze large datasets and make autonomous allocation decisions.