Pengembangan Algoritma Hybrid Metaheuristik Untuk Penentuan Rute Pengiriman Produk Perishable

The decision to dispatch consumers demand has become a strategic and tactical consideration to be solved in an integrated manner. In this study, the problem of determining routing problem take case study of delivery of perishable product. The routes determination should take into account the unique characteristics of perishable products possess. Perishable products continuously decreases quality over their lifetime. The challenge for distributors is how to minimize the cost of delivering perishable products by taking into account the temperature so that it can serve a number of customers within the specified timeframe,The problem of determining the route on delivery is included in the combinatorial optimization problem, thus causing this problem to be complex to be solved by the exact method. On the other hand, metaheuristic methods are increasingly being developed to be applied in the completion of combinatorial optimizations.This research started from mathematical model of perishable product delivery which pay attention to perishability (quality, temperature, quality loss) and time windows. Based on this model, this research develops the route settlement algorithm of delivery of perishable product using metaheuristic, particle swarm optimization. The algorithm development is required because route determination included in discrete issues. In addition, the development of algorithms to improve performance by combining (hybrid) algorithms, nearest neighbor and particle swarm optimization. Experiments were performed on 2 sets of Solomon data. From the experimental results with the metaheuristic hybrid algorithm is able to provide better performance than pure metaheuristik. Although the solution gap produced by these two algorithms is not very significant, but when viewed from the computation time and the number of iterations required to find the best solution, this metaheuristic hybrid algorithm can save an average time of 17 times from pure metaheuristic algorithm.