## A Hardware Genetic Algorithm for the Traveling Salesman Problem on Splash2 (1995)

Venue: | Field-Programmable Logic and Applications |

Citations: | 24 - 2 self |

### BibTeX

@INPROCEEDINGS{Graham95ahardware,

author = {Paul Graham and Brent Nelson},

title = {A Hardware Genetic Algorithm for the Traveling Salesman Problem on Splash2},

booktitle = {Field-Programmable Logic and Applications},

year = {1995},

pages = {352--361},

publisher = {Springer-Verlag}

}

### Years of Citing Articles

### OpenURL

### Abstract

. With the introduction of Splash, Splash 2, PAM, and other reconfigurable computers, a wide variety of algorithms can now be feasibly constructed in hardware. In this paper, we describe the Splash 2 Parallel Genetic Algorithm (SPGA), which is a parallel genetic algorithm for optimizing symmetric traveling salesman problems (TSPs) using Splash 2. Each processor in SPGA consists of four Field Programmable Gate Arrays (FPGAs) and associated memories and was found to perform 6.8 to 10.6 times the speed of equivalent software on a state-of-the-art workstation. Multiple processor SPGA systems, which use up to eight processors, find good TSP solutions much more quickly than single processor and software-based implementations of the genetic algorithm. The four-processor island-parallel SPGA implementation out performed all other SPGA configurations tested. We conclude noting that the described parallel genetic algorithm appears to be a good match for reconfigurable computing machines and that...

### Citations

2827 |
Adaptation in natural and artificial systems
- Holland
- 1975
(Show Context)
Citation Context ...n. Two parallel versions of the algorithm, taking full advantage of the parallel capabilities of Splash 2, are then presented along with results and conclusions. 2 Genetic Algorithms In 1975, Holland =-=[6]-=- developed an optimization technique, based on the process of natural selection and evolution, which he called the genetic algorithm (GA). Follow-on work since that time has shown its usefulness for o... |

108 |
A Genetic Algorithm for Multiprocessor Scheduling
- Hou, Ansari, et al.
(Show Context)
Citation Context ...on work since that time has shown its usefulness for optimization problems requiring the search of large and complex problem spaces from engineering design to combinatorial optimization to control [7]=-=[8]-=-[9]. During its operation the GA maintains a collection (population) of candidate solutions. Associated with each candidate is a fitness or measure of its quality. The algorithm proceeds by selecting ... |

89 |
SPLASH 2
- Arnold, Buell, et al.
- 1992
(Show Context)
Citation Context ...sources and support for linear systolic and MIMD computing models was important for the implementation of SPGA. 1 Introduction The development of reconfigurable computers such as Splash [1], Splash 2 =-=[2]-=-, and PAM [3] makes it possible to quickly create custom hardware implementations of a wide range of algorithms. In some cases, classes of algorithms not previously suitable for ASIC realization can n... |

74 |
Building and using a highly parallel programmable logic array
- Gokhale, Holmes, et al.
- 1991
(Show Context)
Citation Context ...nterconnect resources and support for linear systolic and MIMD computing models was important for the implementation of SPGA. 1 Introduction The development of reconfigurable computers such as Splash =-=[1]-=-, Splash 2 [2], and PAM [3] makes it possible to quickly create custom hardware implementations of a wide range of algorithms. In some cases, classes of algorithms not previously suitable for ASIC rea... |

59 | Introduction to Programmable Active Memories
- Bertin, Roncin, et al.
- 1989
(Show Context)
Citation Context ...upport for linear systolic and MIMD computing models was important for the implementation of SPGA. 1 Introduction The development of reconfigurable computers such as Splash [1], Splash 2 [2], and PAM =-=[3]-=- makes it possible to quickly create custom hardware implementations of a wide range of algorithms. In some cases, classes of algorithms not previously suitable for ASIC realization can now benefit fr... |

28 |
Rapid Implementation of a Genetic Sequence Comparator Using Field Programmable Logic Arrays
- Lopresti
- 1991
(Show Context)
Citation Context ...alization can now benefit from hardware implementation on these machines. Splash 2 and PAM applications have been shown to often run orders of magnitude faster than equivalent software implementations=-=[4]-=-[5]. In this work we describe the result of mapping a genetic algorithm for solving the traveling salesman problem onto the Splash 2 system. We demonstrate the ? This work was supported by ARPA/CSTO u... |

22 |
Parallel genetic algorithms: Introduction and overview of current research
- Dorigo, Maniezzo
- 1993
(Show Context)
Citation Context ...new and better areas of the search space and has been shown to be robust in dealing with a variety of difficult optimization problems due to its tendency to avoid premature convergence to local minima=-=[11]-=-. 3 Splash 2 SPGA was developed on Splash 2, a reconfigurable computer developed at the Center for Computing Sciences (formerly the Supercomputing Research Center). Splash 2 was designed to support li... |

9 |
Programmable active memories: A performance assessment
- Berlin, Roncin, et al.
- 1992
(Show Context)
Citation Context ...zation can now benefit from hardware implementation on these machines. Splash 2 and PAM applications have been shown to often run orders of magnitude faster than equivalent software implementations[4]=-=[5]-=-. In this work we describe the result of mapping a genetic algorithm for solving the traveling salesman problem onto the Splash 2 system. We demonstrate the ? This work was supported by ARPA/CSTO unde... |

9 |
Planning of Component Placement/Insertion Sequence and Feeder Setup in PCB Assembly using Genetic Algorithm
- Leu, Wong, et al.
- 1993
(Show Context)
Citation Context ...thin a sub-tour contained within the original tour. The endpoints of this sub-tour are chosen randomly for each tour mutated. These crossover and mutation operations closely follow those described in =-=[10]-=-. The reader is directed there for details on their operation. The above algorithm readily lends itself to parallelization in at least two ways. The brute force approach is to run multiple independent... |

6 |
Adaptive hydrogenerator governor tuning with a genetic algorithm
- Lansberry, Wozniak
- 1994
(Show Context)
Citation Context ...work since that time has shown its usefulness for optimization problems requiring the search of large and complex problem spaces from engineering design to combinatorial optimization to control [7][8]=-=[9]-=-. During its operation the GA maintains a collection (population) of candidate solutions. Associated with each candidate is a fitness or measure of its quality. The algorithm proceeds by selecting can... |

3 |
Mcm layer assignment using genetic search
- Vemuri, Vemuri
- 1994
(Show Context)
Citation Context ...ow-on work since that time has shown its usefulness for optimization problems requiring the search of large and complex problem spaces from engineering design to combinatorial optimization to control =-=[7]-=-[8][9]. During its operation the GA maintains a collection (population) of candidate solutions. Associated with each candidate is a fitness or measure of its quality. The algorithm proceeds by selecti... |