EVOLUTIONARY ALGORITHMS TO SOLVE LOOSELY CONSTRAINED PERMUT-CSPS: A PRACTITIONERS APPROACH
Luis de‐Marcos, Antonio García‐Cabot, Garcia Eva · International journal of innovative computing, information & control · 2012
Permutation constraint satisfaction problems (permut-CSPs) can be found in many practical applications, wherein most instances usually have a low density of constraints. This paper explores two evolutionary approaches to solve this kind of problem from a practical perspective. A test case that captures the main characteristics present in real world applications is used to design and test the performance of a PSO agent and a GA agent, which are also systematically tuned to determine the best congurations using statistical analysis. We conclude that the PSO agent provides a bettert to the nature of loosely constrained permut-CSPs, resulting in better performance. This paper focuses on the trade-off between development costs (including tuning) and the performance of both evolutionary algorithms, and aims to help practitioners choose the best algorithm and conguration for these problems.