The paper presents an improved hybrid genetic algorithm (GA) forsolving the multiconstrained 0-1 knapsack problem (MKP). Based on thesolution of the LP relaxed MKP, an efficient pre-optimization of theinitial population is suggested. Furthermore, the GA uses sophisticatedrepair and focal improvement operators which are applied to each newlygenerated solution. Care has been taken to define these new operators ina way avoiding problems with the loss of population diversity. The newalgorithm has been empirically compared to other previous approaches byusing a standard set of “large sized” test data. Resultsshow that most of the time the new GA converges much faster to bettersolutions, in particular for large problems
展开▼
