An Incentive Compatible Modification of the Heal Algorithm

Motivated by the study of time-varying resource allocation algorithms, this paper presents the Heal algorithm. We demonstrate how this algorithm can be used to converge to the optimal allocation if each participant cooperatively (and truthfully) follows the rules of the algorithm. However, one drawback of this and other time varying algorithms is that there is a strong incentive for the participants to act noncooperatively and misrepresent, to others, information known only to them. The purpose of this paper is to propose a simple modification to the Heal algorithm which eliminates the incentive for the participants to misrepresent information. The basic ideas used in this paper cab be generalized and are reported elsewhere.