Both AI and robotics are rapidly growing fields in today's society, however a digital mind is not inherently able to navigate the world around it, regardless of if that world is a virtual world or the real world. In order to solve this problem path-finding algorithms were created, these algorithms allow a digital mind to navigate a space as long as it is given enough information to properly calculate what path it should take. However there exists a multitude of different path-finding algorithms and ever more environments that the digital mind could be required to traverse, so which algorithm is the best? Today there is no clear answer, but it is possible to answer the question of: Which of these algorithms perform best in this specific environment? That is what this study aims to do, answer the question of which of following the algorithms, A*, Dijkstra and backtracking performs best in a 3D grid where not every node can be traversed. The metrics we will be using to evaluate the algorithms will be time and memory usage as well as briefly touching upon scalability. In order to gather sufficient data we performed a number of tests on three different grid sizes. The collected metrics shows that the Backtracking algorithm performs the best in 2 out of 3 metrics utilized in the study, however there are complicating factors that require discussion.