Heuristic and metaheuristic techniques are used for solving computationally hard optimization problems. Local search is a heuristic technique while Ant colony optimization (ACO), inspired by the ants' foraging beh...Heuristic and metaheuristic techniques are used for solving computationally hard optimization problems. Local search is a heuristic technique while Ant colony optimization (ACO), inspired by the ants' foraging behavior, is one of the most recent metaheuristic technique. These techniques are used for solving optimization problems. Multiple-Input Multiple-Output (MIMO) detection problem is an NP-hard combinatorial optimization problem. We present heuristic and metaheuristic approaches for symbol detection in multi-input multi-output (MIMO) system. Since symbol detection is an NP-hard problem so ACO is particularly attractive as ACO algorithms are one of the most successful strands of swarm intelligence and are suitable for applications where low complexity and fast convergence is of absolute importance. Maximum Likelihood (ML) detector gives optimal results but it uses exhaustive search technique. We show that 1-Opt and ACO based detector can give near-optimal bit error rate (BER) at much lower complexity levels. Comparison of ACO with another nature inspired technique, Particle Swarm Optimization (PSO) is also discussed. The simulation results suggest that the proposed detectors give an acceptable performance complexity trade-off in comparison with ML and VBLAST detectors.展开更多
文摘Heuristic and metaheuristic techniques are used for solving computationally hard optimization problems. Local search is a heuristic technique while Ant colony optimization (ACO), inspired by the ants' foraging behavior, is one of the most recent metaheuristic technique. These techniques are used for solving optimization problems. Multiple-Input Multiple-Output (MIMO) detection problem is an NP-hard combinatorial optimization problem. We present heuristic and metaheuristic approaches for symbol detection in multi-input multi-output (MIMO) system. Since symbol detection is an NP-hard problem so ACO is particularly attractive as ACO algorithms are one of the most successful strands of swarm intelligence and are suitable for applications where low complexity and fast convergence is of absolute importance. Maximum Likelihood (ML) detector gives optimal results but it uses exhaustive search technique. We show that 1-Opt and ACO based detector can give near-optimal bit error rate (BER) at much lower complexity levels. Comparison of ACO with another nature inspired technique, Particle Swarm Optimization (PSO) is also discussed. The simulation results suggest that the proposed detectors give an acceptable performance complexity trade-off in comparison with ML and VBLAST detectors.