Small-scale Instances Research Articles

A Study of Multi-Constraints Emergency Transportation Problem in Disaster Response

The paper studies a real-world Multi-Constraints Emergency Transportation Problem in disaster response under travel time uncertainty(MCETP), whose objective is to determine the set of flights and routes and the airplane and vehicle assignment for delivering necessary living supplies to disaster areas as quickly as possible. The MCETP can be modeled as a variant of the Vehicle Routing Problem (VRP), which considers a heterogeneous fleet with multi-depots delivering supplies to customers in order to minimize the maximal travel time of all routes where travel time of vehicles is predicted by experts because ground transportation network is destroyed and affected degree is unknown. By adding copy points and imaginary depots, the MCETP can be transformed into a variant of the Travelling Salesman Problem (TSP), for which a Mixed Integer Linear Programming (MILP) formulation is presented. With the proposed initial solution construction algorithm, vehicle assignment strategy and penalty function, the transformed problem can be solved by the extension of Lin–Kernighan–Helsgaun solver (LKH-3). Computational experiments on two sets of totally 50 instances show that the idea of transforming a challenging complex problem to a widely studied problem and using its state-of-the-art algorithm to solve the original problem is highly effective and efficient for the MCETP. Specifically, it can obtain the same solution quality as GUROBI but with much less time for small-scale instances. Furthermore, the proposed algorithm can solve large-scale instances while GUROBI fails to obtain a feasible solution in most cases in a reasonable time.

Two-stream small-scale pedestrian detection network with feature aggregation for drone-view videos

Detecting small-scale pedestrians in aerial images is a challenging task that can be difficult even for humans. Observing that the single image based method cannot achieve robust performance because of the poor visual cues of small instances. Considering that multiple frames may provide more information to detect such difficult case instead of only single frame, we design a novel video based pedestrian detection method with a two-stream network pipeline to fully utilize the temporal and contextual information of a video. An aggregated feature map is proposed to absorb the spatial and temporal information with the help of spatial and temporal sub-networks. To better capture motion information, a more refined flow net (SPyNet) is adopted instead of a simple flownet. In the spatial stream subnetwork, we modified the backbone network structure by increasing the feature map resolution with relatively larger receptive field to make it suitable for small-scale detection. Experimental results based on drone video datasets demonstrate that our approach improves detection accuracy in the case of small-scale instances and reduces false positive detections. By exploiting the temporal information and aggregating the feature maps, our two-stream method improves the detection performance by 8.48% in mean Average Precision (mAP) from that of the basic single stream R-FCN method, and it outperforms the state-of-the-art method by 3.09% on the Okutama Human-action dataset.

Minimizing maximum tardiness on a single machine with family setup times and machine disruption

Due to the complexity of manufacturing processes, many industries today try to group their production components into families and use a common setup time to process every family. Also, the failure of machinery has an important impact on the planning of production systems. This paper deals with the problem of single machine scheduling with family setup times and randomly machine breakdown. It is assumed that there is disruption along the planning horizon with uncertain start and duration times based on a specific probability distribution function. The goal is to minimize the maximum expected tardiness. For this problem, a hill climbing based heuristic (HCO) and a novel hybrid variable neighborhood search metaheuristic (HVNS) have been developed. A branch and bound (BB) procedure is also proposed to optimally solve this problem using HCO and HVNS algorithms to derive upper bounds. Comparison of the solutions obtained from the presented algorithms with the mathematical programming model of the problem in small-scale instances shows that the solutions obtained from the proposed algorithms have a good quality in terms of the expected value of the objective function and the solving time. The computational results for executing the proposed algorithms on 36 series of various instances show that the BB procedure is able to solve 97.76% of the instances within the specified time frame and HVNS metaheuristic is capable to solve 41.07% of them, optimally. Also, the maximum mean deviation of HVNS solutions from the optimal solution is reported below 5% for hard instances and below 15% for instances with medium complexity. After performing a non-parametric hypothesis test, it is found that at a confidence level 99.9%, the deviation index for HVNS is less than 1.5% for difficult problems and less than 9% for semi-difficult ones. Finally, the results of solving large scale instances, demonstrate the efficiency of proposed metaheuristic.

Global-Affine and Local-Specific Generative Adversarial Network for semantic-guided image generation

<p style='text-indent:20px;'>The recent progress in learning image feature representations has opened the way for tasks such as label-to-image or text-to-image synthesis. However, one particular challenge widely observed in existing methods is the difficulty of synthesizing fine-grained textures and small-scale instances. In this paper, we propose a novel Global-Affine and Local-Specific Generative Adversarial Network (GALS-GAN) to explicitly construct global semantic layouts and learn distinct instance-level features. To achieve this, we adopt the graph convolutional network to calculate the instance locations and spatial relationships from scene graphs, which allows our model to obtain the high-fidelity semantic layouts. Also, a local-specific generator, where we introduce the feature filtering mechanism to separately learn semantic maps for different categories, is utilized to disentangle and generate specific visual features. Moreover, we especially apply a weight map predictor to better combine the global and local pathways considering the highly complementary between these two generation sub-networks. Extensive experiments on the COCO-Stuff and Visual Genome datasets demonstrate the superior generation performance of our model against previous methods, our approach is more capable of capturing photo-realistic local characteristics and rendering small-sized entities with more details.</p>

A MATHEMATICAL MODEL AND GENECTIC ALGORITHM SOLUTION METHODS FOR THE BERTH ALLOCATION PROBLEM WITH SEVERAL TYPES OF MACHINES

Maritime shipping is vital to worldwide commerce. Due to the high flow in ports throughout the world, the efficient allocation of vessels in berths has become a problem. A new mathematical model and several algorithms are proposed in this paper to planning the allocation of the vessels in berths and the allocation of resources to the service of each vessel. Those resources, in general, are machines to load or unload vessels. The mathematical model was implemented on Cplex and can solve small scale instances, due to its high complexity. To solve larger instances, a genetic algorithm-based metaheuristic, a first-in first-out heuristic, and a machine allocation algorithm are also proposed in this paper. The model and the algorithms produce very useful and interesting results. Comparing, the results produced by the GA are, on average, 94% better than the results of the Cplex and 26% better than the results of FIFO.

A branch-and-price approach for the parallel drone scheduling vehicle routing problem

This paper considers a delivery system in which a fleet of trucks and unmanned aerial vehicles (or drones) operate independently to serve a set of customers in the minimum possible time. Three mixed integer linear programming (MILP) formulations are proposed for this problem, one of which is an arc-based formulation and the other two are set covering based formulations based on column generation. Owing to the strength of set covering based formulations in producing good linear relaxations, a branch-and-price (BAP) algorithm based on the second set covering based formulation is designed. To solve large-scale problems, a heuristic version of the BAP algorithm is proposed as well. Results from extensive numerical testing indicate that the performances of both the exact and heuristic versions of the BAP algorithm are comparable to the arc-based formulation in solving small-scale instances. Moreover, the heuristic version of the BAP algorithm significantly outperforms the arc-based formulation both in terms of solution quality and runtime for larger problems. Finally, the impact of a UAV power consumption model, according to which, power consumption depends on both speed and payload, is analyzed in the context of this problem. Specifically, a detailed study is performed to show the benefits of optimizing UAV speeds in reducing delivery makespan when UAV endurance is estimated using this power consumption model.

A location-routing problem for biomass supply chains

Facility location and vehicle routing are two critical decision problems in the design of biomass supply chains. Different from previous studies considering these two problems separately, this paper integrates the facility location and the vehicle routing problems for biomass supply chains into a single one, namely the location-routing problem for biomass supply chains (LRP-BSCs). A mixed integer programming model is established for the LRP-BSCs, which can achieve the optimal decisions for small-scale instances. Due to the intrinsic computational complexity of the LRP-BSCs, a hierarchical heuristic algorithm based on Tabu Search is developed for the problem solution. Comprehensive computational examples verify that the proposed approach is effective and efficient.

Consultant assignment and routing problem with priority matching

The consultant assignment and routing problem is to simultaneously assign consultant supplies to client demands and determine the best traveling routes for consultants. Constraints to be considered include skill requirement, capacity limitation, fixed demand, and a maximum number of travel legs. This paper further takes into consideration priority matching, which restricts that clients can only be assigned to consultants with appropriate priority levels. In order to solve this problem, we present a decomposition algorithm named RMIP and a MIP-based neighborhood search algorithm. In addition, we extend an existing MILP formulation and compare our algorithms with it. The effectiveness and efficiency of the proposed algorithms are evaluated on a hundred synthetic instances and twelve real-life instances. Computational results show that the modified MILP formulation is only suitable for solving small-scale instances and a part of the medium-scale instances. For large-scale and real-life instances, the proposed two algorithms are significantly superior to the MILP formulation in solution quality and computational time.

A hybrid multi-level optimisation framework for integrated production scheduling and vehicle routing with flexible departure time

This paper investigates an integrated scheduling problem of production and outbound distribution with flexible vehicle departure time in a time-sensitive make-to-order supply chain. We develop a hybrid multi-level optimisation framework by decomposing the problem into three sub-problems, including vehicle assignment, parallel machines scheduling and distribution scheduling. In this framework, we propose an efficient procedure to obtain the optimal vehicle departure time and utilise metaheuristics and heuristics to obtain the values of other decision variables. Results from extensive numerical experiments indicate that the proposed framework can solve small-scale instances optimally, and for large-scale instances it also shows the better performance than the compared genetic algorithm in terms of convergence and solution quality. Besides, the distribution cost can be reduced by setting flexible vehicle departure time.

Multidepot Heterogeneous Vehicle Routing Problem for a Variety of Hazardous Materials with Risk Analysis

This study investigates a multidepot heterogeneous vehicle routing problem for a variety of hazardous materials with risk analysis, which is a practical problem in the actual industrial field. The objective of the problem is to design a series of routes that minimize the total cost composed of transportation cost, risk cost, and overtime work cost. Comprehensive consideration of factors such as transportation costs, multiple depots, heterogeneous vehicles, risks, and multiple accident scenarios is involved in our study. The problem is defined as a mixed integer programming model. A bidirectional tuning heuristic algorithm and particle swarm optimization algorithm are developed to solve the problem of different scales of instances. Computational results are competitive such that our algorithm can obtain effective results in small-scale instances and show great efficiency in large-scale instances with 70 customers, 30 vehicles, and 3 types of hazardous materials.

Bilevel Optimization for the Hazmat Transportation Problem with Lane Reservation

In this study, we investigate a bilevel optimization model for the hazmat transportation problem with lane reservation. The problem lies in selecting lanes to be reserved in the network and planning paths for hazmat transportation tasks. The trade-off among transportation cost, risk, and impact on the normal traffic is considered. By using the traffic flow theory, we quantify the impact on the normal traffic and modify the traditional risk measurement model. The problem is formulated as a multiobjective bilevel programming model involving the selection of reserved lanes for government and planning paths for hazmat carriers. Two hybrid metaheuristic algorithms based on the particle swarm optimization algorithm and the genetic algorithm, respectively, are proposed to solve the bilevel model. Their performance on small-scale instances is compared with exact solutions based on the enumeration method. Finally, the computational results on large-scale instances are compared and sensitivity analysis on the key parameters is presented. The results indicate the following: (1) Both algorithms are effective methods for solving this problem, and the method based on the particle swarm optimization algorithm requires a shorter computation time, whereas the method based on the genetic algorithm shows more advantages in optimality. (2) The bilevel model can effectively reduce the total risk of the hazmat transportation while considering the interests of hazmat carriers and ordinary travellers. (3) The utilization rate of reserved lanes increases with an increasing number of tasks. Nevertheless, once the proportion of hazmat vehicles becomes excessive, the advantage of reducing the risk of the reserved lanes gradually decreases.

Modeling and Solution Algorithm for Fourth-Party Logistics Routing Problem Based on Service Composition

Aiming at the problems of limited transportation resources and high transportation costs of a single third-party logistics supplier, our research proposes a route planning model based on service combination from the perspective of fourth-party logistics, which aims to find the route with minimum total cost under the premise of meeting customer needs. In order to improve solution efficiency, a specific genetic algorithm is designed. The algorithm adopts backtracking idea and adaptive operations, which can avoid infeasible solution caused by strict constraints, while reducing the time required for crossover and mutation operations to repair infeasible solutions. In addition, Dijkstra-based heuristic algorithm, traditional genetic algorithm and traditional ant colony algorithm are implemented. The comparison of computational experiments on different scales instances shows that the specific genetic algorithm can obtain the optimal solution in small-scale instances, and the solution quality in large-scale problems is better than other algorithms.

Machine learning-driven algorithms for the container relocation problem

The container relocation problem is one of important issues in seaport terminals which could bring a significant saving on the operating cost even with a slight improvement due to the huge number of containers processed across the world each year. Given a specific layout and container retrieval priorities, the container relocation problem aims to find the optimal movement sequence to minimize the total number of container relocation operations. In this paper, we propose novel machine learning-driven algorithms, which integrate optimization methods and machine learning techniques, to solve the problem. More specifically, we propose a new upper bound method called MLUB that incorporates branch pruners. These pruners are derived from some machine learning techniques through using the optimal solution values of many small-scale instances. The tightened upper bounds generated by MLUB are used subsequently both in the exact branch-and-bound algorithm called IB&B and the hybrid beam search heuristic called MLBS. Moreover, we also provide a tighter lower bound for the problem by additionally considering the interaction between consecutive target containers. Based on the benchmark data published recently in the literature, extensive experiments are conducted to test the performance of the proposed algorithms. The experimental results demonstrate that the proposed algorithms outperform the state-of-the-art algorithms reported in the literature, and some managerial insights regarding the load intensity of the bay and some algorithm parameters such as the look-ahead depth and the beam width are drawn from the results.

Towards delivery-as-a-service: Effective neighborhood search strategies for integrated delivery optimization of E-commerce and static O2O parcels

In this paper, we investigate a new variant of last-mile delivery that integrates the scheduling of static E-commerce parcels and Online-to-Offline(O2O) parcels. The O2O parcels, such as flowers and cakes, are often delivered intra city with a time window constraint. It is driven by the concept of delivery-as-a-service, which targets at building consolidated infrastructure and using the same fleet of vehicles to provide standardized delivery services for different types of merchants. We formulate it as an integration of Multi-Depot Multi-Trip Vehicle Routing Problem (MDMTVRP) and Paired Pickup and Delivery Problem with Time Window (PPDPTW). To solve the mixed problem of MDMTVRP and PPDPTW, we present its Mixed-Integer Programming (MIP) model to obtain the optimal solution for small-scale instances. To solve large-scale problems, we propose a hybrid neighborhood search strategy to effectively combine the merits of ALNS and tabu search. We also present a two-level pruning strategy that can significantly accelerate the local search procedure. We conduct extensive numeric experiments on multiple datasets, and results showed that our hybrid approach achieved near-optimal performance and established clear superiority over ALNS and tabu search.

Effective algorithms for single-machine learning-effect scheduling to minimize completion-time-based criteria with release dates

Multi-variety and small-batch productions are usually undertaken by skilled workers instead of an automatic assembly line because of economic cost consideration. In the production process, a worker's familiarity to an operation influences the length of task execution time. An interesting phenomenon called learning effect has become a trending research topic. This study investigates a learning effect scheduling model on a single machine system, in which the learning effect is position-dependent and each task is released at different dates. Two optimal criteria are individually discussed: one is total k-power completion time, and the other is maximum lateness. Both problems are NP-hard, therefore, effective algorithms are provided to handle different scale problems within an appropriate CPU time. The heuristic algorithms, namely, shortest processing time available and earliest due date available, are introduced to achieve feasible schedules for large-scale instances, and their asymptotic optimality is proven given that the problem scale tends to infinity. The two heuristics can thus serve as optimal algorithms in mass production. For small-scale instances, a branch and bound algorithm is presented to achieve the optimal solution, where a release-date-based branching rule and preemption-based lower bounds eliminate as many invalid nodes as possible. For medium-scale instances, an evolutionary-based metaheuristic algorithm, namely, discrete differential evolution, is utilized to seek high-quality solutions, in which the initial population and crossover operator are well-designed to enhance its performance. A number of random experiments demonstrate the superiority of the proposed algorithms.

Exact and heuristic algorithms for scheduling on two identical machines with early work maximization

Early work criterion is a new-proposed objective in the scheduling field, in which the early parts of jobs processed before their due dates are aimed to be maximized. In this paper, we consider such a criterion for scheduling on two identical machines. For the simplified model with a common due date of all jobs, we show that the classical LPT (Largest Processing Time first) heuristic is an approximation algorithm with the worst case ratio 109. For the general model with arbitrary due dates, we propose a branch-and-bound algorithm equipped with several truncation rules, upper and lower bounds techniques, as well as dominance rules to solve small scale instances, while a genetic algorithm to solve big instances. Finally, the efficiencies of the proposed algorithms are validated by the numerical experiments, in which we show that the branch-and-bound algorithm can find the exact solutions within a reasonable time-consumption when an instance is small, and the genetic algorithm is an efficient approach when the problem instance turns to big.

Partial disassembly line balancing under uncertainty: robust optimisation models and an improved migrating birds optimisation algorithm

A partial disassembly line balancing problem under uncertainty is studied in this paper, which concerns the allocation of a sequence of tasks to workstations such that the overall profit is maximised. We consider the processing time uncertainty and develop robust solutions to accommodate it. The problem is formulated as a non-linear robust integer program, which is then converted into an equivalent linear program. Due to the intractability of such problems, the exact algorithms are only applicable to small-scale instances. We develop an improved migrating birds optimisation algorithm. Two enhancement techniques are proposed. The first one finds the optimal number of tasks to be performed for each sequence rather than random selection used in the literature; while the second one exploits the specific problem structure to construct effective neighbourhoods. The numerical results show the strong performance of our proposal compared to CPLEX and the improved gravitational search algorithm (IGSA), especially for large-scale problems. Moreover, the enhancement due to the proposed techniques is obvious across all instances considered.

Bus tour-based routing and truck deployment for small-package shipping companies

This study investigates a joint optimisation of routing and truck deployment for a small-package shipping company. Bus tour-based services are provided to collect express packages at customer bases. The objective of the optimisation problem is to minimise the average waiting time for packages stored at the customer bases. We first propose a mixed-integer nonlinear programming model. The proposed formulation is linearised and made solvable by an off-the-shelf mixed-integer linear programming solver. For solving larger instances of the problem, we propose two optimisation approaches that can obtain near-optimal solutions – a local branching-based method and a particle swarm optimisation-based method. We conduct numerical experiments to validate the effectiveness and efficiency of the proposed solution methods. The results indicate that both methods can obtain optimal solutions quickly for most of the small-scale instances. For medium-size instances, the local branching-based method performs best, and the PSO-based method outperforms the others for large-size instances.

An energy-efficient permutation flowshop scheduling problem

The permutation flowshop scheduling problem (PFSP) has been extensively explored in scheduling literature because it has many real-world industrial implementations. In some studies, multiple objectives related to production efficiency have been considered simultaneously. However, studies that consider energy consumption and environmental impacts are very rare in a multi-objective setting. In this work, we studied two contradictory objectives, namely, total flowtime and total energy consumption (TEC) in a green permutation flowshop environment, in which the machines can be operated at varying speed levels corresponding to different energy consumption values. A bi-objective mixed-integer programming model formulation was developed for the problem using a speed-scaling framework. To address the conflicting objectives of minimizing TEC and total flowtime, the augmented epsilon-constraint approach was employed to obtain Pareto-optimal solutions. We obtained near approximations for the Pareto-optimal frontiers of small-scale problems using a very small epsilon level. Furthermore, the mathematical model was run with a time limit to find sets of non-dominated solutions for large instances. As the problem was NP-hard, two effective multi-objective iterated greedy algorithms and a multi-objective variable block insertion heuristic were also proposed for the problem as well as a novel construction heuristic for initial solution generation. The performance of the developed heuristic algorithms was assessed on well-known benchmark problems in terms of various quality measures. Initially, the performance of the algorithms was evaluated on small-scale instances using Pareto-optimal solutions. Then, it was shown that the developed algorithms are tremendously effective for solving large instances in comparison to time-limited model.

A framework for preemptive multi-skilled project scheduling problem with time-of-use energy tariffs

The growing importance of energy consumption has become an integral part of many decision-making processes in various organizations. In this paper, a framework is proposed for an organization where the undertaken project is implemented by multi-skilled workforce and energy tariffs depending on the time-of-use. This is a real-life situation where energy tariffs are significantly high during the peak demand compared to the one in off peak demand in order to control overloaded consumption of energy. A formulation is developed for the problem to minimize the total cost of consuming required energy. The proposed formulation is validated using several small-scale instances solved by GAMS software. To tackle large-scale instances, two intelligent meta-heuristics called electromagnetic like algorithm (EMA) and genetic algorithm (GA) are utilized. Furthermore, a comprehensive analysis is performed based on 50 test instances to evaluate the intelligence and the robustness of the employed algorithms. Finally, statistical tests are used to compare the performances of the utilized EMA and GA.