The problem of mutual exclusion has to be solved to prevent race condition and, as a result, prevent the possibility of a program producing an incorrect result. Providing deadlock-free distributed mutual exclusion algorithms is often difficult and it involves passing many messages. The two major types of these algorithms are token-based and permission-based algorithms. In this research, we propose a hybrid distributed mutual exclusion algorithm. By Hybrid, we mean that the algorithm uses both token-based and permission-based techniques. The best case and worst case number of messages passed for every critical region entry and exit is calculated, which are better than many other algorithms.

Keywords