References of "Arda, Yasemin"
     in
Bookmark and Share    
Full Text
Peer Reviewed
See detailMulti-period vehicle loading with stochastic release dates
Arda, Yasemin ULg; Crama, Yves ULg; Kronus, David ULg et al

in EURO Journal on Transportation and Logistics (2014), 3(2), 93-119

This paper investigates a multi-period vehicle loading problem with stochastic information regarding the release dates of items to be transported. The deterministic version of the problem can be ... [more ▼]

This paper investigates a multi-period vehicle loading problem with stochastic information regarding the release dates of items to be transported. The deterministic version of the problem can be formulated as a large-scale set covering problem. Several heuristic algorithms are proposed to generate decision policies for the stochastic optimization model over a long rolling horizon. The resulting policies have been extensively tested on instances which display the main characteristics of the industrial case-study that motivated the research. The tests demonstrate the benefits of the multi-period stochastic model over simple myopic strategies. A simple and efficient heuristic is shown to deliver good policies and to be robust against errors in the estimation of the probability distribution of the release dates. [less ▲]

Detailed reference viewed: 157 (35 ULg)
Full Text
See detailOptimization of the service start time for an elementary shortest path problem with time windows
Arda, Yasemin ULg; Crama, Yves ULg; Kucukaydin, Hande ULg

E-print/Working paper (2014)

We investigate an elementary shortest path problem with resource constraints where a single capacitated vehicle, initially located at a depot, must serve a set of customers while respecting their ... [more ▼]

We investigate an elementary shortest path problem with resource constraints where a single capacitated vehicle, initially located at a depot, must serve a set of customers while respecting their individual time windows. When the vehicle visits a customer, it delivers the customer's demand and collects a revenue in return for the delivery. The vehicle can start its trip at any desired time. The transportation cost is a function of both the total distance traveled and the duration of the assigned trip. The objective is to determine the service start time from the depot, the subset of customers to be served, and the trip to be performed so as to minimize the total loss, which is calculated as the di erence between the transportation cost and the revenue collected from the customers. We develop two exact dynamic programming algorithms which can deal with an in nite number of Pareto-optimal states arising from the fact that the starting time and the duration of the trip act like continuous decision variables. We report computational results obtained with these algorithms and with a faster heuristic for the elementary shortest path problem. We also examine the performance of these algorithms when they are used to solve the pricing subproblem arising in the framework of a column generation algorithm for a related vehicle routing problem with time windows. [less ▲]

Detailed reference viewed: 69 (6 ULg)
See detailVehicle routing problems with multiple trips: using specific local search operators
Arda, Yasemin ULg; Crama, Yves ULg; François, Véronique ULg et al

Conference (2014, June 25)

In vehicle routing problems with multiple trips (VRPM), each vehicle is allowed to perform more than one trip during its working period. Classical solution techniques for this problem make use of existing ... [more ▼]

In vehicle routing problems with multiple trips (VRPM), each vehicle is allowed to perform more than one trip during its working period. Classical solution techniques for this problem make use of existing VRP heuristics to create trips, together with bin packing methods aimed at assigning these trips to the available vehicles. The first contribution of this work is to propose specific local search operators for the VRPM. The operators directly integrate the multi-trip structure of the problem within well-known VRP operators. As a second contribution, heuristics using these operators are compared with classical solution techniques mentioned above. The comparison is performed by using the adaptive large neighborhood search metaheuristic as a common basis for both methods. The most classical version of the problem is studied as well as a variant involving time windows [less ▲]

Detailed reference viewed: 46 (6 ULg)
Peer Reviewed
See detailAn exact bi-directional dynamic programming algorithm for an elementary shortest path problem with variable service start time
Arda, Yasemin ULg; Crama, Yves ULg; Kucukaydin, Hande ULg

Conference (2014, January 31)

We consider an elementary shortest path problem with resource constraints (ESPPRC), where a capacitated single vehicle serves customers by respecting their associated time windows. The vehicle can start ... [more ▼]

We consider an elementary shortest path problem with resource constraints (ESPPRC), where a capacitated single vehicle serves customers by respecting their associated time windows. The vehicle can start servicing the customers at any desired time, but it can be used for a fixed amount of time. The total transportation cost depends on the total distance traveled and the total amount of time that the vehicle spends by performing the assigned trip. On the other hand, each served customer yields a revenue. Thus, the aim is to identify the path to be followed and the start time of the vehicle from the depot that minimize the total transportation cost minus the gained revenues. This kind of a problem can be encountered as a pricing subproblem when a branch-and-price algorithm is applied to solve vehicle routing problems with additional constraints. In such a case, the revenues correspond to the dual prices of the visited vertices. It is known that the classical ESPPRC can be solved to optimality by implementing a dynamic programming (DP) algorithm. However, our problem has to take an infinite number of Pareto-optimal states into consideration, since the vehicle can leave the depot at any point in time and charges depending on the total traveling and waiting time. We propose an exact DP algorithm which can deal with an infinite number of Pareto-optimal states by representing total traveling and waiting time as a piecewise linear function of the service start time at the depot and develop suitable dominance rules. Furthermore, a column generation algorithm is devised for solving the relaxed set covering formulation of the related vehicle routing problem where new columns are determined by the proposed DP algorithm. Finally, computational results are presented. [less ▲]

Detailed reference viewed: 43 (0 ULg)
Peer Reviewed
See detailOptimization of Service Start Time for an Elementary Shortest Path Problem
Kucukaydin, Hande ULg; Arda, Yasemin ULg; Crama, Yves ULg

Conference (2013, July 08)

We are concerned with an elementary shortest path problem with resource constraints (ESPPRC), where there is a capacitated single vehicle at the depot for serving a set of delivery and backhaul customers ... [more ▼]

We are concerned with an elementary shortest path problem with resource constraints (ESPPRC), where there is a capacitated single vehicle at the depot for serving a set of delivery and backhaul customers with a time window. On a given route, the vehicle can visit a backhaul customer only after all its delivery customers are visited, where the delivery and backhaul customers are considered to be two disjoint sets. Split deliveries and pick-ups are not allowed. In this problem, the vehicle may be assigned to several routes. In addition, the vehicle can begin servicing the customers at any desired time and can be used for at most a fixed amount of time that depends on the shift duration of the assigned driver. Distance and time based variable costs are incurred by serving the customers. Namely, the total cost depends on the total distance traveled and the total amount of time that the vehicle spends by performing the assigned multiple trips. On the other hand, serving a customer yields also a revenue. Therefore, the objective is to determine the optimal service start time of the vehicle from the depot along with the trips to be performed in order to minimize the total of the distance and time costs minus the collected revenues. Such a problem can be faced as the pricing subproblem in branch-and-price algorithms for vehicle routing problems with additional constraints, where the revenues are equivalent to the dual prizes of the visited vertices. In general, ESPPRC can be solved to optimality by using a dynamic programming algorithm. However, since the vehicle can start the service at any point in time and is paid based on the total time during which it has been used, our ESPPRC has to take an infinite number of Pareto-optimal states into account. Therefore, we adapt the well-known dynamic programming algorithm according to this feature and develop piecewise linear time functions that represent total traveling and waiting time depending on a variable start time at the depot. Consequently, we propose appropriate dominance rules to discard feasible paths that cannot lead to the optimal solution. Finally, computational results are presented. [less ▲]

Detailed reference viewed: 57 (11 ULg)
See detailA two-stage stochastic programming model for a comprehensive disaster management problem
Döyen, Alper ULg; Arda, Yasemin ULg

Conference (2013, July 03)

A periodic disaster preparedness model that incorporates mitigation, response, and recovery related decisions is proposed. The objective is to minimize the total cost of relief item transportation and ... [more ▼]

A periodic disaster preparedness model that incorporates mitigation, response, and recovery related decisions is proposed. The objective is to minimize the total cost of relief item transportation and shortage as well as recovery costs of buildings and links under a limited retrofitting budget. The problem is formulated as a two-stage stochastic programming model. While retrofitting decisions are the binary first-stage variables, relief flows and shortage amounts constitute continuous second-stage variables. An efficient solution approach that employs the integer L-shaped method is proposed. [less ▲]

Detailed reference viewed: 30 (4 ULg)
Full Text
See detailGeneral lotsizing in a closed-loop supply chain with uncertain returns
Amand, Guillaume ULg; Arda, Yasemin ULg

Conference (2013, July 03)

We consider a multi-product capacitated lotsizing and scheduling prob- lem with sequence-dependent setups and stochastic product returns. The returned products accumulate in an input inventory and can be ... [more ▼]

We consider a multi-product capacitated lotsizing and scheduling prob- lem with sequence-dependent setups and stochastic product returns. The returned products accumulate in an input inventory and can be sold as new items after a remanufacturing process. The determinis- tic demand of end items can also be satisfied through a manufacturing process that is fed by an unlimited source of raw materials. An ap- proximate dynamic algorithm is developed to solve both single-item and multi-items cases. [less ▲]

Detailed reference viewed: 29 (3 ULg)
Peer Reviewed
See detailAn elementary shortest path problem with variable service start time
Kucukaydin, Hande ULg; Arda, Yasemin ULg; Crama, Yves ULg

Conference (2013, May 02)

We consider an elementary shortest path problem with resource constraints (ESPPRC), where a capacitated single vehicle serves a set of delivery and backhaul customers with a revenue and a time window. On ... [more ▼]

We consider an elementary shortest path problem with resource constraints (ESPPRC), where a capacitated single vehicle serves a set of delivery and backhaul customers with a revenue and a time window. On a given route, the vehicle can visit a backhaul customer only after all its delivery customers are visited, where a customer is either a delivery or a backhaul customer, but not both. Split deliveries and pick-ups are not allowed. In this problem, multiple trips are allowed so that the vehicle can be assigned to several routes. In addition, the vehicle can begin servicing the customers at any desired time and can be used for at most a fixed amount of time that depends on the shift duration of the assigned driver. Distance and time based variable costs are incurred by serving the customers. In other words, the total cost depends on the total distance traveled and the total amount of time that the vehicle spends by performing the assigned multiple trips. On the other hand, serving a customer yields also a revenue. Therefore, the aim is to determine the optimal service start time of the vehicle from the depot along with the trips to be performed in order to minimize the routing costs minus the collected revenues. Such a problem can be faced as the pricing subproblem in branch-and-price algorithms for vehicle routing problems with additional constraints, where the revenues are equivalent to the dual prizes of the visited vertices. In general, ESPPRC can be solved to optimality by using a dynamic programming algorithm. However, since the vehicle can start the service at any point in time and is paid based on the total time during which it has been used, our ESPPRC has to take an infinite number of Pareto-optimal states into account. Therefore, we adapt the well-known dynamic programming algorithm according to this feature and develop piecewise linear time functions that represent total traveling and waiting time depending on a variable start time at the depot. Consequently, we propose appropriate dominance rules to discard feasible paths that cannot lead to the optimal solution. Finally, we present computational results. [less ▲]

Detailed reference viewed: 15 (1 ULg)
Full Text
See detailAn adaptive large neighborhood search for a vehicle routing problem with multiple trips and driver shifts
Arda, Yasemin ULg; Crama, Yves ULg; François, Véronique ULg

Conference (2013, February 13)

This study analyzes a rich vehicle routing problem with multiple trips and driver shifts. The considered problem features are inspired from the practical case of a Belgian distribution company. Along with ... [more ▼]

This study analyzes a rich vehicle routing problem with multiple trips and driver shifts. The considered problem features are inspired from the practical case of a Belgian distribution company. Along with the multi-trip component, characteristics of this particular problem include time windows, pickup and delivery customers, and site-vehicle dependencies. Internal and external fleets are considered with different cost structures and driver shifts constraints. An adpative large neighborhood search is used to treat the problem. [less ▲]

Detailed reference viewed: 176 (8 ULg)
See detailTwo-stage stochastic integer programming models for strategic disaster preparedness
Döyen, Alper ULg; Aras, Necati; Arda, Yasemin ULg

Conference (2013, February 08)

We are interested in two distinct disaster management models which are modeled by two-stage stochastic integer programming. While the first model solution provides optimal post-disaster response related ... [more ▼]

We are interested in two distinct disaster management models which are modeled by two-stage stochastic integer programming. While the first model solution provides optimal post-disaster response related decisions and pre-disaster retrofitting decisions to minimize the total cost, in addition to the first model, the second model considers the recovery actions and the time of earthquake occurrence. In the models, retrofitting decisions are given for both buildings and bridges under a limited budget. Effective and efficient solution methods are proposed in the study. [less ▲]

Detailed reference viewed: 25 (4 ULg)
Full Text
See detailA general lot-sizing problem in a closed-loop supply chain with returns
Amand, Guillaume ULg; Arda, Yasemin ULg

Conference (2013, February 08)

We consider a stochastic version of the multi-product multi-level capacitated lotsizing and scheduling problem with sequence-dependent setups. A bottler needs to determine its production schedule over a ... [more ▼]

We consider a stochastic version of the multi-product multi-level capacitated lotsizing and scheduling problem with sequence-dependent setups. A bottler needs to determine its production schedule over a finite horizon in order to satisfy a deterministic demand. The raw materials are supplied through two different sources: uncapacitated reserves of new bottles and the uncertain returns of used bottles. We present results for the single-item case. [less ▲]

Detailed reference viewed: 75 (2 ULg)
Full Text
Peer Reviewed
See detailMultiperiod vehicle loading with stochastic release dates
Arda, Yasemin ULg; Crama, Yves ULg; Kronus, David ULg et al

Conference (2012, May 24)

Production scheduling and vehicle routing problems are well-known topics in operations management. Although these tasks are consecutive in the supply chain, few optimization models tackle the associated ... [more ▼]

Production scheduling and vehicle routing problems are well-known topics in operations management. Although these tasks are consecutive in the supply chain, few optimization models tackle the associated issues. A most common situation, in practice, is actually that transportation management is disconnected from production planning: when production items or batches have been completely processed by the manufacturing plant, they become available for shipping, and they are consequently handled by the transportation managers. From a global managerial perspective, and with a view towards coordination of the product flows and customer satisfaction, this is not an ideal process. It is by far preferable, indeed, to set up an integrated production-transportation plan taking into account, among other constraints, the capacity of the plants and the customer due-dates. The present research proposes a methodology to investigate a multi-period vehicle loading problem with deterministic or stochastic information concerning items arrivals from production. Results from related optimization techniques are statistically compared and the benefits of the multi-period and stochastic modeling is demonstrated. Finally, an efficient heuristic is highlighted and is shown to be robust to the deviation from item arrival forecasts. [less ▲]

Detailed reference viewed: 8 (3 ULg)
Full Text
Peer Reviewed
See detailA Rich Vehicle Routing Problem with Multiple Trips and Driver Shifts
Arda, Yasemin ULg; Crama, Yves ULg; Kucukaydin, Hande ULg et al

Conference (2012)

This study is concerned with a rich vehicle routing problem (RVRP) encountered at a Belgian transportation company in charge of servicing supermarkets and hypermarkets belonging to a franchise. The ... [more ▼]

This study is concerned with a rich vehicle routing problem (RVRP) encountered at a Belgian transportation company in charge of servicing supermarkets and hypermarkets belonging to a franchise. The studied problem can be classified as a one-to-many-to-one pick-up and delivery problem, where there is a single depot from which all delivery customers are served and to which every pick-up demand must be carried back (Gutiérrez-Jarpa et al., 2010). The delivery and backhaul customers are considered to be two disjoint sets, where on a given route backhaul customers can be visited only after all delivery customers are served. Split deliveries and pick-ups are not allowed. The service at a customer must start within the given time window of the customer. However, it is not possible to serve a customer with every available vehicle, since the vehicles at the company’s disposal are of different types according to which the capacity changes. Therefore, our problem can be classified also as a heterogeneous fleet vehicle routing problem with customer-vehicle incompatibilities (Ceselli et al., 2009). The problem at hand requires that the same vehicle may be assigned to several routes, which leads to a multiple-trip RVRP. Furthermore, driver shifts are taken into account so that each vehicle of the fleet starts servicing the customers when the shift of the assigned driver starts. The shift duration is the same for all drivers. If the service of the vehicle exceeds SH, an overtime cost incurs to the transportation company. In such a case, a vehicle can be used during at most a fixed length of time. In addition to the vehicles of the company’s own fleet (i.e. internal vehicles), there is a possibility to request external vehicles for servicing some customers. External vehicles can be used for a fixed maximum amount of time and can start servicing customers at any desired time. A fixed reservation cost and distance and time based variable costs are incurred in the case of an external vehicle, while only distance based variable costs are incurred in the case of an internal vehicle. We employ a binary integer linear programming formulation in order to model our problem. The first constraint set ensures that each customer is visited at least once by either an internal or external vehicle. With the second constraint set, it is guaranteed that each internal vehicle is assigned to at most one tour. The last two constraint sets are the binary restrictions on the assignment variables. In order to solve the problem, we first relax the binary restrictions on the assignment variables and develop a column generation procedure, where we obtain two pricing problems, one for internal vehicles and the other one for external vehicles. The pricing problem of each internal vehicle can be formulated as an elementary shortest path problem with resource constraints, which can be solved using a dynamic programming algorithm based on a bounded bi-directional search (Righini and Salani, 2008). However, since an external vehicle can start the service at any point in time and is paid based on its total travel time, the second pricing algorithm has to take into account an infinite number of Pareto-optimal states (Liberatore et al., 2011). We discuss the efficient solution of the pricing subproblems and present preliminary computational results. [less ▲]

Detailed reference viewed: 186 (5 ULg)
Full Text
See detailA capacity game in transportation management
Amand, Guillaume ULg; Arda, Yasemin ULg

E-print/Working paper (2012)

Detailed reference viewed: 32 (16 ULg)
Full Text
See detailA capacity game in transportation management
Amand, Guillaume ULg; Arda, Yasemin ULg

Conference (2011, July 10)

This study analyzes the contractual relation between a retailer and a carrier with the aim of determining possible deviations from the optimal system performance. In order to face a random demand, the ... [more ▼]

This study analyzes the contractual relation between a retailer and a carrier with the aim of determining possible deviations from the optimal system performance. In order to face a random demand, the retailer submits a contract to the carrier based on the number of units transported and on the number of truck used. Then, before uncertainty is resolved, the carrier decides of the number of trucks that he reserves. Once the demand is known, the carrier may also request additional trucks at a higher cost. The result shows that the proposed contract does not coordinate the supply chain. [less ▲]

Detailed reference viewed: 25 (11 ULg)
Full Text
See detailMultiperiod vehicle loading optimization with stochastic supply
Amand, Guillaume ULg; Arda, Yasemin ULg; Crama, Yves ULg et al

Scientific conference (2011, April 07)

Detailed reference viewed: 16 (6 ULg)
Full Text
Peer Reviewed
See detailA capacity game in transportation management
Amand, Guillaume ULg; Arda, Yasemin ULg

Conference (2011, March 03)

Emerging concerns about competitiveness induce a growing number of firms to outsource their outbound transportation operations to third-party logistics providers. The resulting increase in the number of ... [more ▼]

Emerging concerns about competitiveness induce a growing number of firms to outsource their outbound transportation operations to third-party logistics providers. The resulting increase in the number of actors often leads to sub-optimal supply chain actions due to the antagonistic nature of the economic objectives of the partners. With the aim of determining possible deviations from the optimal system performance in such supply chains, this study analyzes the contractual relation between a retailer and a third-party logistics provider (carrier) using game theoretical approaches. The partners of the studied supply chain play a Stackelberg game in which the retailer is the leader and the carrier is the follower. The retailer faces an uncertain demand and needs to supply his store from his warehouse. he has the option of not meeting all the demand but must satisfy at least a minimum proportion of the fi nall demand. On the other hand, the carrier has to determine the number of trucks needed to satisfy this demand before uncertainty is resolved. Once demand is realized, if the reserved transportation capacity is insufficient, the carrier also has the possibility to requisition trucks at a higher price. We modelise the problem and propose a contract having two parameters : the quantity of transported items and the number of truck used. In our settings, the retailer is the one that submits the contract and the carrier decides if he accepts it or not. We compare this situation with a centralized model where a single decision maker takes all the decisions. [less ▲]

Detailed reference viewed: 75 (27 ULg)