You are on page 1of 5

Modules 1. ACO 2. FCS-ANTMINE 3.

Weighted FCS-ANTMINER

Description 1. ACO ACO-based algorithms are stochastic search procedures that inspired from the behavior of real ant colonies. Real ants are able to find the shortest path from a food source to their nest. They deposit a chemical substance called pheromone on the ground while walking Pheromone helps them to communicate with one another Ants can smell pheromone and when choosing their way, they tend to choose, in a probabilistic way, paths marked by more pheromone. If there is no pheromone in the environment, ants select paths in a random manner. Initially, there is no pheromone on the ground and therefore, ants choose the paths arbitrarily. Pheromone is evaporated over time. Therefore, in shorter paths pheromone evaporation is less than longer paths in which fewer numbe. This positive feedback means that since more pheromone stimulates more number of ants therefore all the ants will eventually select the shortest path. Fig. 1 shows how ants can find the shortest path. Recently, the ACO meta-heuristic was proposed as a common framework for existing applications. Each ant builds a possible solution to the problem by moving through a finite sequence of neighbor states (nodes). Moves are selected by applying a stochastic local search directed by the ant internal state, problemspecific local information and the shared information about the pheromone.

Edge selection An ant is a simple computational agent in the ant colony optimization algorithm. It iteratively constructs a solution for the problem at hand. The intermediate solutions are referred to as solution states. At each iteration of the algorithm, each ant moves from a state to state , corresponding to a more complete intermediate solution. Thus, each

ant

computes a set

of feasible expansions to its current state in each iteration, of moving from values, viz.,

and moves to one of these in probability. For ant , the probability state to state depends on the combination of two

the attractiveness

of the move, as computed by some heuristic indicating the a of the move, indicating how

priori desirability of that move and the trail level

proficient it has been in the past to make that particular move. The trail level represents a posteriori indication of the desirability of that move. Trails are updated usually when all ants have completed their solution, increasing or decreasing the level of trails corresponding to moves that were part of "good" or "bad" solutions, respectively. In general, the th ant moves from state to state with probability

where is the amount of pheromone deposited for transition from state is a parameter to control the influence of (a priori knowledge, typically . , , where and to , 0

is the desirability of state is the distance) and 1

transition

is a parameter to control the influence of

represent the attractiveness

and trail level for the other possible state transitions. Pheromone update When all the ants have completed a solution, the trails are updated by where

is the amount of pheromone deposited for a state transition the pheromone evaporation coefficient and

is

is the amount of pheromone deposited

by th ant, typically given for a TSP problem (with moves corresponding to arcs of the graph) by

where

is the cost of the th ant's tour (typically length) and

is a constant.

2. FCS-ANTMINE The ants compete with one another to make the best rules. But increasing the competition causes some ants to select the high accuracy rules while other ants are unable to find proper paths to choose high quality rules. Therefore, the output rule set contents both high and low accuracy rules and despite some high quality rules, the final classification system may be unable to classify significant number of input patterns. The main reason for this problem is the lack of enough cooperation among the ants. That is, rather than increasing the total system accuracy, each ant makes an effort to learn individual rules with high accuracy. Although pheromone helps them to have cooperation but it is not enough due to their high and disruptive competition. In other hand, if we do not control the cooperation of the ants, it might lead the algorithm to be trapped in local optimum. Therefore, it seems essential to define a strategy to balance cooperation and competition of the ants

At first iteration (t = 0), ant0 makes rule Rh and in the next iterations (tP1) the ants modify rule Rh. The maximum terms that each ant can modify, in each iteration (tP1), is determined by a parameter named Max_change. This parameter can be the number of attributes at most. Let termij be a rule condition of the form (Ti is Aj), where Ti is the ith attribute and Aj is the jth antecedent fuzzy set (Fig. 2). Each ant chooses termi,j to modify (or make rule Rh in the first iteration) according to the following probability:

where gi,j is a problem-dependent heuristic value for termij. Higher values of gi,j, increases the probability of termi,j to be chosen.

3. Weighted FCS-ANTMINER

Weighted fuzzy is same as the FCS-ANTIMINER, but the only change is we will be assigning separate weights for all parameters. For rule classification, all the parameter will involve in selecting a class. If any parameter has more dependency for the classification the we will assign more weight to that particular parameter. So that it will increase the classification accuracy.

You might also like