An improved genetic algorithm for the multiconstrained 0-1 knapsack problem

The paper presents an improved hybrid genetic algorithm (GA) for solving the multiconstrained 0-1 knapsack problem (MKP). Based on the solution of the LP relaxed MKP, an efficient pre-optimization of the initial population is suggested. Furthermore, the GA uses sophisticated repair and focal improvement operators which are applied to each newly generated solution. Care has been taken to define these new operators in a way avoiding problems with the loss of population diversity. The new algorithm has been empirically compared to other previous approaches by using a standard set of “large sized” test data. Results show that most of the time the new GA converges much faster to better solutions, in particular for large problems