A peer to peer ridesharing system connects drivers who are using their personal vehicles to conduct their daily activities with passengers who are looking for rides. A well-designed and properly implemented ridesharing system can bring about social benefits, such as alleviating congestion and its adverse environmental impacts, as well as personal benefits in terms of shorter travel times and or financial savings for the individuals involved. In this paper, the goal is to study the impact of availability of carpool lanes and traffic conditions on ridesharing demand using an agent based simulation model. Agents will be given the option to use their personal vehicles, or participate in a ridesharing system. An exact many to many ride matching algorithm, where each driver can pick up and drop off multiple passengers and each passenger can complete his or her trip by transferring between multiple vehicle, is used to match drivers with passengers. The proposed approach is implemented in AnyLogic ABS software with a real travel data set of Los Angeles, California. The results of this research will shed light on the types of urban settings that will be more recipient towards ridesharing services.