A Domain Theory for Task Oriented Negotiation

We present a general theory tha t captures the re la t ionship between certain domains and nego t ia t ion mechanisms. T h e analysis makes i t possible to categorize precisely the k inds of domains in wh ich agents f ind themselves, and to use the category to choose appropr ia te negot i a t ion mechanisms. T h e theory presented here bo th generalizes previous results, and al lows agent designers to characterize new domains accurately. The analysis thus serves as a cr i t ica l step in using the theory of negot ia t ion in realwor ld appl icat ions. We show tha t in certain Task Or iented Domains, there exist d is t r ibu ted consensus mechanisms w i t h s imple and stable strategies t ha t lead to efficient outcomes, even when agents have incomplete i n f o rma t i on about their environment . We also present add i t i ona l novel resul ts, in par t i cu la r tha t in concave domains using a l l -o r -no th ing deals, no l y ing by an agent can be benef ic ial , and t ha t in subaddi t ive domains, there often exist beneficial decoy lies tha t do not require fu l l i n f o rma t i on regarding the other agent's goals.