Paper Title
Using Heuristic Method to Recommendation Systems

Abstract
A heuristic method is applied to recommendation systems. In the problem a lot of objects is needed to classify based on their properties into a positive or a negative set, taking into account the preferences of the ranking person. An agent is created for the ranking process which executes a reordering operation in the produced environment after a selection. The activity of the agent was monitored by the environment and if the measure of reordering reaches a certain limit, the agent gets reward, otherwise, it is penalized. Considering this, the agent attempts to maximize the reward and hereby the ranking is carried out as quickly as possible. The objects (for example motion pictures or products) can be reached and selected on a user interface. At the start there is no object classified. If an object is selected from the initial (the third) set as acceptable, then relying upon this finding the arranging of objects is modified, so the number of elements of the three sets changes. If another object classified, for example again into a set of positive objects, then our algorithm continues the ranking process. With ranking of only some objects can be created a high precision recommendation considering the user preferences. The results suggest that this approach can provide a new method for solving ranking problems with high-performance. Keywords - Agent, Heuristic Method, Recommendation System, Reward.