Find path to a goal

Creates a numerical matrix of coordinates which contain the path points that the agent will use to move towards a goal. It thus finds itself on the strategic level (when performed in an initial planning phase) and the tactical level (when performed in response to blockages).

Usage

find_path(object, ...)

# S4 method for class 'goal'
find_path(
  object,
  agent,
  background,
  space_between = radius(agent),
  many_nodes = FALSE,
  algorithm = "Dijkstra",
  precomputed_edges = NULL,
  new_objects = NULL,
  reevaluate = FALSE
)

Arguments

object: Object of goal-class for which the path should be computed.
...: Arguments passed on to the methods of this generic
agent: Object of agent-class who wants to move towards the goal defined in argument object.
background: Object of background-class within which the goal and agent are contained.
space_between: Numeric denoting the amount of space to leave between an object and a path point. Within this function, the default is the radius of the agent that is provided. However, in the simulate function, this default is overwritten to be equal to 2.5 times this radius.
many_nodes: Logical denoting whether to create many nodes or leave it at the minimum. Defaults to FALSE.
algorithm: Character denoting the algorithm to be used for optimizing the path towards the goal. Is provided to the get_path_pair function in the cppRouting package. Defaults to "Dijkstra". Note that one-way routing as controlled by the limited_access slot in the background may not work if a bidirectional algorithm is used.
precomputed_edges: List containing the network of nodes and edges for the environment, as created by compute_edges. Defaults to NULL, triggering the creation of these edges again.
new_objects: List of instances of object-class that are not yet contained in the objects slot in the argument background. Usually consists of other agents that the agent has to account for when planning their route. Defaults to NULL.
reevaluate: Logical denoting whether to reevaluate the network that is provided in precomputed_edges. Is useful whenever new_objects is not NULL, allowing us to check whether some nodes and edges are now occluded by the new objects. Defaults to FALSE.

Value

Numerical matrix of coordinates representing the path points the agent will have to move to to reach the goal.

Details

Creates a numerical matrix containing coordinates of path points which, together, form a strategic path to the goal specified in this function. This forms one of the functions that operate on the strategic and tactical level of the model.

To create this path, the function takes the following approach. First, it will create a network of nodes that are connected through unidirectional edges, as created by the create_edges function. Nodes are created based on the objects in the environment, so that the nodes are always strategically placed at a certain distance away from the objects in the environment (at a distance space_between). If the argument many_nodes = TRUE, then a grid of other nodes are added to this list of initial nodes. This grid consists of nodes that are placed at an equal distance apart in the x- and y-direction, and spans 20 rows and 20 columns. Once the nodes have been generated, those nodes that fall within any of the objects in the environment or falls outside of the environment are deleted. Unidirectional edges connect the different nodes and are similarly pruned based on intersections with objects or, when defined, intersections with instances of the segment-class.

Once the network of nodes and edges has been created, we translate this network into a graph that can be used by the ccpRouting package. Specifically, we use makegraph for this purpose. We then use the get_path_pair function of the cppRouting package to find the shortest route to the goal. The user can specify which algorithm to use for this optimization through the argument algorithm, but the algorithm should be unidirectional to ensure that one-way walking can be enforced (as controlled through the limited_access slot in the background-class).

Finally, the optimal path is checked for any redundancy and formatted so that the path points are contained in an n x 2 matrix.

Examples

# Create a setting
my_background <- background(shape = rectangle(center = c(0, 0), 
                                              size = c(2, 2)),
                            objects = list(circle(center = c(0, 0), 
                                                  radius = 0.5)))

# Create an agent that is walking around there and a goal that the agent 
# will move to
my_goal <- add_goal(objects(my_background)[[1]], my_background)
my_agent <- agent(center = c(0.7, 0.7), radius = 0.25)

# Find the path of the agent to his goal
find_path(my_goal, 
          my_agent,
          my_background)
#>              x         y
#> [1,] 0.3169533 0.3995505

Usage

Arguments

Value

Details

See also

Examples