Method of distributed aggregation in a call center
View Patent ↗A method for partitioning a call center having N agents associated with M agent groups, for computation by a plurality of computational nodes, has steps for (a) assigning each agent as a vertex in a hypergraph; (b) assigning each agent group as a hyper-edge in the hypergraph; and (c) applying a hypergraph partitioning algorithm to partition the agents and groups relative to the nodes with the hypergraph cost function awarding equal load distribution to nodes and minimizing inter-node traffic.
1. A method for partitioning a call center having agents associated with agent groups, for computation by a plurality of computational nodes, comprising:
(a) assigning, by a processor, the agents as a vertex in a hypergraph;
(b) assigning, by the processor, the agent groups as a hyper-edge in the hypergraph; and
(c) applying, by the processor, a hypergraph partitioning algorithm configured to partition the agents and the agent groups relative to the computational nodes with a hypergraph cost function awarding equal load distribution to the computational nodes;
(d) mapping, by the processor, the agents and the agent groups as an object to one of the computational nodes by a function, such that each computational node is an owner for the object;
(e) for the agents, computing, by the processor, agent aggregates locally by an associated owner node of the computational nodes; and
(f) for the agent groups, computing, by the processor, agent group aggregates for each agent group shared by the computational nodes by:
i) creating a partial state record of the agent group aggregates by each computational node by aggregating on the agents owned by each computational node and belonging to each agent group; and
ii) sending the partial state records to the owner node, which combines the partial state records into a total state record,
wherein the partitioning of the call center is for minimizing data traffic between the computational nodes.
2. The method of claim 1 , further comprising assigning, by the processor, a weight to each vertex, the vertex corresponding to the agents.
3. The method of claim 2 , where the weight is based on a computational complexity of the agent aggregates of the agents.
4. The method of claim 1 , further comprising partitioning, by the processor, the hypergraph into partitions corresponding to the computational nodes.
5. The method of claim 4 , further comprising using, by the processor, the partitioning to assign each of the agents and the agent groups to a corresponding one of the computational nodes.
6. The method of claim 4 , further comprising providing the partitioning in real time.
7. The method of claim 1 , where the function comprises a Hash function.
8. The method of claim 1 , where a hypergraph cost function minimizes inter-node traffic between the computational nodes.
9. A system for partitioning a call center having agents associated with agent groups, for computation by a plurality of computational nodes, comprising:
a processor; and
a non-transitory physical medium, wherein the medium has instructions stored thereon that, when executed by the processor, causes the processor to:
assign the agents as a vertex in a hypergraph;
assign the agent groups as a hyper-edge in the hypergraph; and
apply a hypergraph partitioning algorithm configured to partition the agents and the agent groups relative to the computational nodes with a hypergraph cost function awarding equal load distribution to the computational nodes;
map the agents and the agent groups as an object to one of the computational nodes by a function, such that each computational node is an owner for the object;
compute, for the agents, agent aggregates locally by an associated owner node of the computational nodes; and
compute, for the agent groups, agent group aggregates for each agent group shared by the computational nodes by:
creating a partial state record of the agent group aggregates by each computational node by aggregating on the agents owned by each computational node and belonging to each agent group; and
sending the partial state records to the owner node, which combines the partial state records into a total state record,
wherein the partitioning of the call center is for minimizing data traffic between the computational nodes.
10. The system of claim 9 , further comprising assigning, by the processor, a weight to each vertex, the vertex corresponding to the agents.
11. The system of claim 10 where the weight is based on a computational complexity of the agent aggregates of the agents.
12. The system of claim 9 , further comprising partitioning, by the processor, the hypergraph into partitions corresponding to the computational nodes.
13. The system of claim 12 , further comprising using, by the processor, the partitioning to assign each of the agents and the agent groups to a corresponding one of the computational nodes.
14. The system of claim 12 , further comprising providing the partitioning in real time.
15. The system of claim 9 , where the function comprises a Hash function.
16. The system of claim 9 , where a hypergraph cost function minimizes inter-node traffic.