strategies.txt

用C编写的部落寻优的优化算法

Strategies to use (cf. problems.txt)

Normal problem: 12 12 12 1
or			: 18 18 18 1
Difficult problem: 21 21 21 21
(Partly) combinatorial problem: 21 21 21 21



case 0: // Around p_i OR around p_g
case 1: // Mixture "around p_i" "around p_g". The best compromise
case 2: // Mixture "around x" "around p_i" "around p_g"
case 3: // Try to find promising areas (center and radius) and choose one at random
case 4: // Mixture "around p_i" "around p_g" and "particle previous move"
case 5: // Mixture "around p_i" "around p_g" and "particle previous move"

case 9: // "Classical PSO". Mixture "around p_i" "around p_g" and "particle previous move"
            // with hyperparallelepids
case 10:  // Keep moving. Neighbours are not taken into account.
case 11:   // Go back.   Neighbours are not taken into account.
case 12: // Mixture "around p_i" "around p_g" + noise on position
case 13: // As 1, noise on function evaluation will be added after
case 14: // Mixture "around p_i" "around p_g" , using Direct Gaussian distribution
case 15: // Mixture "around p_i" "around p_g" , "around x", using Direct Gaussian distribution (pivot method)
case 16: // Mixture "around p_i" "around p_g" , "around x",
            //pivot method (using Direct Gaussian distribution if nonunif<0)
case 17:  // Simplified mixture (x,p_i,p_g)
case 18:  // Pivot method with hyperspheres
case 19: // Pivot method with mixture "around p_i" "around pivot"
case 20: // Semi-hypersphere in direction x(t-1)->x(t), centered on x
case 21: // Hyperparallelepids, for problems with non null granularity
			// Gaussian on each dimension
case 22: // Hyperparallelepids, for problems with non null granularity
			// Intervals