Modeling R&D Strategy as a Network Search Problem
1) Node search: breadth-first, search-only.
If all visited nodes have been acquired, then identify the parents of the acquired nodes. Sort those nodes in order of the cost of visiting them from a previously acquired node. Visit as many of the desired nodes for sale as can be afforded within the technology search budget of the agent.
2) Node search: breadth-first, willing-to-buy, purchase-node.
If all visited nodes have been acquired, then identify which among the parents of the acquired nodes are available for sale by other agents. Sort those which are for sale in order of purchase price. Starting with the cheapest, buy as many of the desired nodes for sale as can be afforded within the technology search budget of the agent.
3) Node search: breadth-first, willing-to-buy, search-for-node.
If all visited nodes have been acquired, then identify which among the parents of the acquired nodes are not available for sale by other agents. Sort those which are not for sale in order of the cost of visiting them from a previously acquired node. Visit as many of the desired nodes for sale as can be afforded within the technology search budget of the agent after deducting the cost of any node purchases.
4) Node search: depth-first, search-only.
If all visited nodes have been acquired, then identify the parents of the acquired nodes and visit the parent node which is the cheapest in the sense that the arc between that node and any acquired node has the label of smallest value among the arcs to any of the acquired nodes.
5) Node search: depth-first, willing-to-buy, purchase-node.
If all visited nodes have been acquired, then identify which among the parents of the acquired nodes are available for sale by other agents. Sort those which are for sale in order of purchase price. Buy the cheapest of the desired nodes for sale that can be afforded within the technology search budget of the agent.
6) Node search: depth-first, willing-to-buy, search-for-node.
If all visited nodes have been acquired and all of the parents of the acquired nodes that are for sale have already been purchased, then sort the parent nodes which are not for sale in order of the cost of visiting them from a previously acquired node. Visit the cheapest of the desired nodes for sale that can be afforded within the technology search budget of the agent after all purchases of nodes.
7) Child node search: search-only.
If any nodes have been visited but not acquired, then sort the children of such nodes in order of the values of the arcs from visited nodes to their children. Starting with the cheapest, visit as many of the desired child nodes as can be afforded within the technology search budget of the agent.
8) Child node search: willing-to-buy, purchase-node.
If any nodes have been visited but not acquired and are for sale by other agents, then sort such nodes in order of the values of the labels on the arcs from visited nodes. Starting with the cheapest, buy as many of the desired nodes for sale as can be afforded within the technology search budget of the agent.
9) Child node search: willing-to-buy, search-for-node.
If the parents of one or more acquired nodes have been visited but not themselves acquired, then sort the unacquired children of such nodes in order of the values of the labels on the arcs from the parents to the children. Starting with the cheapest, visit as many of the desired child nodes as can be afforded within the technology search budget of the agent after all purchases of nodes. (This rule ensures that agents willing to buy in technology will only search one level down from nodes they have identified as being among those they want to move to from their acquired technology. They will not undertake a full development project going back down the technology network. When the node becomes available for purchase, then that option will always be taken.)
Generated with CERN WebMaker