[[alternative]]Using Genetic Algorithms to control Elevator Group System

[[abstract]]The problem of Elevator Group Control System (EGCS) is how to schedule elevators smartly and continuously in a dynamic and real-time environment. Because it is a concurrent system, slightly differences in system scheduling can produce different results. For example, same test cases may yield different results in average riding time, average waiting time, or crowdedness of passengers. Such phenomenon poses great challenge to EGCS. Because there are too many factors must take into consideration by EGCS, we use Genetic algorithm to control the elevators of EGCS, which has been known as an effective approach to find near-optimal solutions. We implement an elevator simulation system using Java’s multi-threading constructs. On which, we experiment two methods: (1) using Genetic algorithm to schedule the elevators. (2) using shortest distance to schedule the elevators. We compare the results and conclude that the approach using Genetic algorithm yields better results.