High level deal-making can be an arduous task, so negotiators often bring in a mediator. But there are problems with human mediators: you never know for sure whether they're unbiased, whether they might reveal confidential information by accident, or whether they're fully competent to mediate the negotiation.
But there's hope: operations researchers are now championing computerized mediators as the solution. The commercial version is a few years off, but the US and other governments are already using mecha-mediators on a limited basis. OR/MS Today explains how it works:
- Both sides come up with their own "value function," which is an equation that defines the minimum acceptable terms of any potential agreement.
- The value functions are entered into the computer, and some nifty software determines whether they intersect.
- If the functions don't intersect, the parties can break off their talks, or they can lower their minimum demands. Knowing up front whether there's room to move can save a huge amount of time wasted in dead-end negotiation.
- If the functions do intersect, then there's room to negotiate. That's good news, because it immediately shows that there's at least one acceptable agreement.
- Now the two parties can have the computer optimize the deal in a utilitarian way, where the total gain from the deal is maximized. (Think mergers here, where both sides want the new company's value to be maximized).
- Or, the negotiators can proceed through the happy task of splitting up the excess value themselves, with the computer there analyzing value for them.
All those calculations require some pretty high-level math programming, but it's easy to envision a commercial software tool that would function in much the same way. Let's make a deal!
