agalbachicar / platform_analytics Goto Github PK

The simulator evolves the state of the graph and agents. It may interact with the controller.

The procedure to evolve the time follows:

• for agent in agents:
  • grabs the adjacent node to traverse
    • if adjacent node has capacity: moves the agent to next node
    • else: keep the node in current node
    • time in task ++
• time without task ++

Given the set of nodes in a graph, it should create random tasks that would involve two nodes. Once two nodes are used, they are blocked until they are repositioned again.

Tasks are given a rate lambda.

Create a wharehouse graph that can be parametrized:

• Each node has:
  • Name: n_i_j
  • Capacity: non-negative integer that tells how many robots can stay indefinitely there. It does not block flow.
  • Available capacity: varies with t based on how many agents stay in the node.
• Each arc has:
  • Name: n_i_j_n_k_l
  • Capacity: non-negative integer that tells how many robots can flow simultaneously.
  • Available capacity: varies with t based on how many agents flow through in the node.

Layout will be rectangular. In the sides and in the middle, vertical arcs have capacity 4, the rest 2.

The controller has a list of robots and knows if they are assigned or not.

The controller is responsible for assigning tasks to agents based on their cost but knowing the congestion in the zone.

In

An agent is robot.

It moves with velocity v through an arc e between two nodes. v is a node property. Agents will be called r_i where i belongs to [0; R-1].

An agent can compute a trajectory given a graph that allows him to move from node A to node B. The trajectory has a cost measured in time.

An agent will be queried to go from A to B, stay certain time in B and then go to C. An agent will compute the time it takes, a trajectory composed of a list of pairs <node, time in node> from A to B to C.

An agent could be assigned or not to a task.

An agent can release a node to be used again for another task after spending a certain amount of time there.