An Effective Order-Aware Hybrid Genetic Algorithm for Capacitated Vehicle Routing Problems in Internet of Things

Abstract

The Internet of Things integrates lots of capacitated vehicles in smart logistics. The routing for capacitated vehicles is a combinatorial optimization problem which has been widely studied in recent years. This paper proposes an effective order-aware hybrid genetic algorithm for the capacitated vehicle routing problem in the Internet of Things. The method is characterized by an improved initialization strategy and a problem-specific crossover operator. The former combines the sweep algorithm with randomness to harmonize the contradiction between diversity and convergence, while the latter integrates neighborhood search heuristics to find the offspring with the best fitness and check constraints simultaneously. A large number of simulations have been carried out, and the results validated the effectiveness of our algorithm. INDEX TERMS Internet of Things, capacitated vehicle routing problem, order-aware hybrid genetic algorithm, sweep algorithm, neighborhood search heuristics.

Operations management of smart logistics: A literature review and future research

“…For instance, the Optimization Cold chain logistics Decentralized and centralized carbon trading systems provide a reasonable revenue-and-cost-sharing contract with the trade-off among the market demand, total carbon emission, and supplier's profit Ma et al (2020) Warehouse management An IoT-based warehouse management system maximizes warehouse productivity and picking accuracy with computational intelligence techniques Lee et al (2018) Routing optimization User-centric logistics service model using ontology: Users' smart devices are applied to collect the location and situation information from the delivery vehicle and user for routing optimization IoT devices security Blockchain-based security system framework: Multiple-agreement algorithm is applied to enable thin-plate spline performance, which is efficient in solving the vulnerability of IoT multi-platform security Kim et al (2018) Authentication of object tracking Fine-grained IoT-enabled object tracking system: Cloud storage, symmetric key cryptosystem, and one-way hash function are applied to perform end-to-end authentication protocol Anandhi et al (2019) multi-object recognition algorithm improved by Chen et al (2016) integrates the slotted anti-collision algorithm with the simulation results of RFID collisions, which deals with the dynamics scanning data of tags. Lin et al (2019) studied the routing for capacitated vehicles in IoT, which is a widely examined combinatorial optimization problem in smart logistics. They proposed a novel order-aware hybrid generic algorithm composed of an improved initialization strategy and a problem-specific crossover operator.…”

Optimization Model and Algorithm for Dockless Bike-Sharing Systems Considering Unusable Bikes in China

“…Each customer can be served only once by only one vehicle. To the best of our knowledge, there are a variety of optimal and heuristic algorithms to MVRPSDP problems in the existing literature [24]- [26], [37]- [39]. In this study, we apply the state-ofthe-art MIP resolver tool, ILOG-CPLEX (version 12.80) to efficiently resolve the relaxed model with Constraints (4)- (10) and (12) and obtain the optimal routes for vehicles.…”

“…The replenishment of usable bikes is seldom considered, which is another critical decision in the maintenance of shared-bikes, since the shortage of usable bikes at hotspots may have a negative impact on travelers' satisfaction and the performance of the system [17], [23]. Besides, the proposed approaches in above literature just reduce reverse logistic to a deterministic multiple vehicle routing problems (MVRP) to pick up unusable bikes within a region without considering the random demands and returns of bikes [24]- [26].…”

A Decision Support Tool for Urban Freight Transport Planning Based on a Multi-Objective Evolutionary Algorithm

“…The meta-heuristics that have been applied in the UFT context are ant colony optimization [30], artificial bee colony [34], biogeography-based algorithms [23], the firefly algorithm [19], evolutionary algorithms [15], [17], [18], [24], [25], [31], [21], taboo search [33], [38] and simulated annealing [29], [40], among others. From the articles that address multi-objective versions of VRP with more than 20 nodes (clients), in more than 80% of them use, as said, meta-heuristics, and among them, in particular Evolutionary Multi-Objective Algorithms (MOEAs) [36].…”