本篇論文針對飯店訂房的配置問題,提出一個混合型和弦搜尋演算法 (Hybrid Harmony Search Algorithm, HHSA),其目的是要在有限的計算時間內找出一組最佳配置解,使得飯店訂房問題獲得最大利潤。混合型和弦搜尋演算法結合了 改良型和弦搜尋演算法 (IHS)與新型 和弦搜尋衍生演算法(NDHS)的優點主要核心參數為和弦記憶大小 (HMS)、和弦記憶考量率(HMCR)、節距調整率(PAR)與距離帶寬(BW),以此為母體進行新可行解之搜尋與組合。接著將所提出的混合型和弦搜尋演算法,應用在飯店訂房限制收益最佳化問題以測試效能,飯店訂房限制收益最佳化問題是屬於困難的隨機模擬最佳化問題,具有很大的解空間。最後將所提出的演算法與粒子群演算法(PSO)、演化式策略(ES)、基因演算法(GA)、模擬退火法(SA)、人工蜂群演算法(ABC)以及傳統的和弦搜尋演算法(HS)等六種演算法進行比較,由模擬數據顯示所提出的混合型和弦搜尋演算法,不論在解的品質和計算效率上,都能獲得很好的測試結果。
In this thesis, a hybrid of harmony search algorithm, abbreviated as HHSA, is proposed to solve the hotel booking limits problem. The goal is to search for a good enough solution with the objective of maximizing the expected revenue using limited computation time. The proposed HHSA utilizes the advantage of improved harmony search algorithm and novel derivative harmony search algorithm. The core parameters contain the harmony memory size, harmony memory considering rate, pitch adjustment rate and distance bandwidth, which are used to search for the optimal solution from entire solution space. Then the proposed HHSA is applied to a hotel booking limits problem, which is formulated as a hard stochastic simulation optimization problem that consists of a huge solution space comprised by the vector of booking limits. Finally, the proposed HHSA is compared with the particle swarm optimization algorithm, evolutionary strategies, genetic algorithm, simulated annealing, artificial bee colony and traditional harmony search algorithm. The vector of good enough booking limits obtained by the proposed HHSA is promising in the aspects of solution quality and computational efficiency.