Zor zaman pencerelerine sahip iki kademeli yer seçimi–rotalama problemlerine evrimsel bir algoritma

Rapid growth in freight distribution networks due to increasing demand created the necessity for effective and efficient methods for freight vehicle movements. Motivated by the effective distribution network design problems, a two-echelon location routing problem with hard time windows (2E-LRPTW) is studied. This problem combines two NP-Hard problems, including strategic and tactical decisions: the facility location problem (FLP) and the vehicle routing problem (VRP). In this study, the first echelon consists of city distribution centers (CDC) and satellites; the second echelon is constituted of interaction between satellites and customers. The network is connected through two types of vehicle fleets with different characteristics. Each type of vehicle leaves the corresponding facility during working hours and returns to it. Imposing capacity restrictions to both facilities and vehicles and adding hard time window constraints to customers, the problem complexity increases. Consequently, an evolutionary algorithm (EA) inspired by a genetic algorithm is proposed to solve large-size instances with good quality within a reasonable time. The EA decides which facilities to open, allocations, and resulting routes originated from each facility at both echelons. Computational experiments and results indicate the proposed EA capable of finding optimal solutions and improving the best-known solutions for some instances.