In this paper, we address the problem of seeking a source that emits signal described by a function that is radially symmetric, and decays with increasing distance in a complex environment with obstacles. Electromagnetic signals, acoustic signals, vapor emission, etc, are examples of such signals. In contradistinction to existing techniques, we use a non-gradient based technique known as Particle Swarm Optimization (PSO) to overcome the difficulties posed due to lack of a mathematical model for the decay profile in real scenarios. We propose static and dynamic obstacle avoidance strategies, and integrate them with PSO in the source seeking problem. Finally, we validate the effectiveness of our strategies with simulation and experiments.