Using Anti-pheromone to Identify Core Objects for Multidimensional Knapsack Problems: A Two-step Ants based Approach

This paper proposes a two-step ants algorithm for the Multidimensional Knapsack Problem. In the first step, the algorithm uses an Anti-pheromone to detect which objects are less suitable to be part of a near-optimal solution solving the opposite problem. From this information, in the second step an ant-based algorithm continues searching for better solutions trying to solve the real problem.

[1]  Thomas Stützle,et al.  MAX-MIN Ant System , 2000, Future Gener. Comput. Syst..

[2]  Marcus Randall,et al.  Anti-pheromone as a Tool for Better Exploration of Search Space , 2002, Ant Algorithms.

[3]  Hamid R. Tizhoosh,et al.  Applying Opposition-Based Ideas to the Ant Colony System , 2007, 2007 IEEE Swarm Intelligence Symposium.

[4]  Günther R. Raidl,et al.  The Core Concept for the Multidimensional Knapsack Problem , 2006, EvoCOP.