A new concept is proposed for an elevator group control systemwhich can change control settings according to individual floorutilization situations. The floor attribute based control method uses acombination of floor attribute based evaluation and car attribute basedevaluation. Computer simulations show it can improve multiplepreferential floor simulations at the same time. Additionally, anon-line parameter tuning method using genetic algorithms is proposed.The floor attribute control method needs some dozens of controlparameters to be tuned according to each building's traffic, and theproposed method can do it. The elitest preserving strategy is aneffective way to keep control stability. A diversity recoveringoperation, which initializes population, excluding the elitestindividual when traffic flow data have changed, improves the followingability for transitions of utilized conditions
展开▼