Wonders of Seabed: Difficulty Evaluation of Management Games Using Neural Network - Technologies and Applications of Artificial Intelligence - page 166

have enough money to pay tax. Subsequently, the population of the city would
decrease as citizens choose to leave the city.
Maintenance Fee. The maintenance fee is the expense that is regularly in-
curred to maintain the city. For example, if there is sucient amount of main-
tenance fee, the city area would be clean. The maintenance fee is used to hire
cleaners. If players allocate the maintenance fee below a certain threshold, the
number of cleaners would become fewer and fewer. Consequently, the city areas
would become dirty, resulting in a 'messy' city. This messy event would lead to
the deterioration of the environment. The total income is decreased due to this
messy event. The government unit would need to raise money to restore the city
environment.
Celebration Free. A celebration event is an event held by the government
unit. The purpose of this event is to give a chance for citizens to relax and also
encourage them to work hard. If the government unit spends sucient amount
of money to hold parties, citizens would be happy, thereby leading to double the
income.
Human Resource Free. The factor of human resource is about public affairs
that are managed by the government unit. If there is less human resource, an
event called 'wrong policy' would happen. Players should make sure that human
resource is kept above a reasonable standard. When a 'wrong policy' occurs, a
quarter of the income would be lost.
3.2 Game Implementation
This section presents the implementation of the game. We build the game as a
3D game. There are agents, fish, 3D models of different buildings. Our game runs
at real time rate. Players can interact with the game via mouse and keyboard
controls. We apply different techniques to control crowd simulation, fish school
simulation, and game balance. Figures 1 shows the objects used in the four areas.
In the part of crowd simulation, we apply the A* algorithm to compute the
movement of crowd. At the beginning of the game, we record all the possible
paths which are computed by using the A* algorithm. Thus, the agents (or
citizens) can be animated on the paths based on their current positions and
destinations, e.g., walking to farms and returning houses. To avoid collision be-
tween agents, our system dynamically adjusts their movement direction if they
would collide with each other. After the collision events are resolved, the agents
continue to move based on the pre-computed paths.
To simulate a school of fish, we compute the distance between fish. We also
compute the distance between each fish and a group of fish. The fish reaction is
performed according to the distance between fish and the group [11]. Figure 2
shows the fish reaction behaviors. Basically, we apply repulsive force to push
away the approaching fish for avoiding collision. We ignore the details about the
mathematical models about controlling the orientation and movement of agents
and fish in this paper.