| 摘要 |
A method and a corresponding computer-readable medium are provided for optimizing a decision assignment based on a sought benefit. The optimization operations include mapping agents to actors into pairs, designating a benefit to each pair for a set of nodes, arranging the nodes into heaps (with each heap corresponding to an agent), selecting the node with the sought benefit as the head of the heap for all the heaps, and summing each benefit of the heads to establish a cumulative benefit. The benefit designation further includes associating the node with the benefit, action and agent that correspond to that pair, and disposing the node into the heap that corresponds to the agent. Arranging the heap further includes comparing first and second nodes to determine which has the sought benefit within the heap, and ordering the peak node as the head of the heap. Further operations include deconflicting first and second heaps that have heads with equal benefit, and truncating tail nodes from the head of each heap. |
