Diversification in crowd simulations

This is my Disertation project from my final year of university entitled: "

Investigating the effects of diversification amongst agents on crowd simulation using a Leader follower relationship" 

the goal of my research was to determine wether or not  introducting agents with different behaviours in this case a leader follower relationship(LFR) would impact the behaviours of the rest of the crowd. my results suggested that there was some impact  however it was mostly neglegible  with the number of agents used (between 500 and 3000), though in the 139 tests run there where consistently changes in congestion accross the map though ususally only by small ammount typically no more than 0.2%. though it is also worth noting that an increase in the number of LFR agents showed an increase in computational complexity.

I have made a modified version of my project and the write up avalible here