Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approa... more Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approach for simulating stochastic transient networks of bulk queues that often arise in consolidation networks. A set of approximations was proposed in order to produce a computationally tractable algorithm. This paper describes the numerical methods actually used for calculating probabilities and the results of extensive numerical experiments which test the accuracy of those approximations and the overall efficiency of the procedure, vis-à ...
ABSTRACT Abstract We address the problem of combining a cost-based simulation model, which makes ... more ABSTRACT Abstract We address the problem of combining a cost-based simulation model, which makes decisions over time by minimizing a cost model, and rule-based policies, where a knowledgeable user would like certain types of decisions to happen with a specified frequency when averaged over the entire simulation. These rules are designed to capture issues that are dicult,to quantify as costs, but which produce more realistic behaviors in the judgment of a knowledgeable user. We consider patterns that are specified as averages over time, which have to be enforced in a model that makes decisions while stepping through time (for example, while optimizing the assignment of resources to tasks). We show how an existing simulation, as long as it uses a cost-based optimization model while stepping through time, can be modified to more closely match exogenously specified patterns.
In this paper we develop a new method for solving Dynamic Resource Allocation Problems DRAP that ... more In this paper we develop a new method for solving Dynamic Resource Allocation Problems DRAP that occur in the operation of freight transportation systems. Such problems involve the allocation of resources to perform tasks over a discrete-time, dynamic network. Our particular interest is in ultra-large scale problems, that involve managing thousands of resources such as drivers or vehicles. We focus on problems with dynamic attributes, where the characteristics of the resource may change as it handles each task. We provide a general and very exible formulation, and provide a solution based on the principle of dynamic programming. A newly proposed linearization approximation is shown to provide high quality solutions with reasonable execution times. The technique is illustrated in the context of the driver management problem for a large motor carrier.
Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived fr... more Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived from an application in less-than-truckload trucking, as a dynamic resource allocation problem. Drivers are aggregated into groups characterized by an attribute vector which capture the important attributes required to incorporate the work rules. The problem is very large: over 5,000 drivers and 30,000 loads in a four-day planning horizon. We formulate a problem that we call the heterogeneous resource allocation problem, which is more ...
Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approa... more Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approach for simulating stochastic transient networks of bulk queues that often arise in consolidation networks. A set of approximations was proposed in order to produce a computationally tractable algorithm. This paper describes the numerical methods actually used for calculating probabilities and the results of extensive numerical experiments which test the accuracy of those approximations and the overall efficiency of the procedure, vis-à ...
Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005., 2005
The allocation of human and physical resources over time is a fundamental problem that is central... more The allocation of human and physical resources over time is a fundamental problem that is central to management science. In this paper we review a mathematical model of dynamic resource allocation that is motivated by problems in transportation and logistics. In principle, problems of this type can be solved via dynamic programming. However, three "curses of dimensionality" give rise to intractable computational requirements. We present computationally efficient approximate dynamic programming algorithms developed by the first author and collaborators for application to problems in freight transportation. We discuss how these algorithms address the three curses of dimensionality and how they relate to other independent threads of research on mathematical programming and approximate dynamic programming.
EURO Journal on Transportation and Logistics, 2012
Deterministic optimization has enjoyed a rich place in transportation and logistics, where it rep... more Deterministic optimization has enjoyed a rich place in transportation and logistics, where it represents a mature field with established modeling and algorithmic strategies. By contrast, sequential stochastic optimization models (dynamic programs) have been plagued by the lack of a common modeling framework, and by algorithmic strategies that just do not seem to scale to real-world problems in transportation. This paper is designed as a tutorial of the modeling and algorithmic framework of approximate dynamic programming, however our perspective on approximate dynamic programming is relatively new, and the approach is new to the transportation research community. We present a simple yet precise modeling framework that makes it possible to integrate most algorithmic strategies into four fundamental classes of policies, the design of which represent approximate solutions to these dynamic programs. The paper then uses problems in transportation and logistics to indicate settings in which each of the four classes of policies represent a natural solution strategy, highlighting the fact that the design of effective policies for these complex problems will remain an exciting area of research for many years. Along the way, we provide a link between dynamic programming, stochastic programming and stochastic search.
2013 46th Hawaii International Conference on System Sciences, 2013
ABSTRACT Given the importance of handling high levels of uncertainty from renewables in the unit ... more ABSTRACT Given the importance of handling high levels of uncertainty from renewables in the unit commitment problem, there has been increased attention given to the use of stochastic programming methods. Since these are computationally very demanding, there is a need for new approximations. We propose to use a point forecast of energy from wind and loads, where the point forecast is chosen and adapted by simulation to produce a robust solution. Traditionally, point forecasts represent an expectation. In our work, we suggest that we can use an appropriately chosen quantile of the forecast distribution which is optimized within a stochastic environment. The result is a policy search algorithm built around a point forecast, which is easily implement able using standard industry models and algorithms.
Proceedings of the 2010 Winter Simulation Conference, 2010
We describe an adaptation of the knowledge gradient, originally developed for discrete ranking an... more We describe an adaptation of the knowledge gradient, originally developed for discrete ranking and selection problems, to the problem of calibrating continuous parameters for the purpose of tuning a simulator. The knowledge gradient for continuous parameters uses a continuous approximation of the expected value of a single measurement to guide the choice of where to collect information next. We show how to find the parameter setting that maximizes the expected value of a measurement by optimizing a continuous but nonconcave surface. We compare the method to sequential kriging for a series of test surfaces, and then demonstrate its performance in the calibration of an expensive industrial simulator.
The Spherical Frontier DEA Model (SFM) (Avellar et al., 2007) was developed to be used when one w... more The Spherical Frontier DEA Model (SFM) (Avellar et al., 2007) was developed to be used when one wants to fairly distribute a new and fixed input to a group of Decision Making Units (DMU's). SFM's basic idea is to distribute this new and fixed input in such a way that every DMU will be placed on an efficiency frontier with a spherical shape. We use SFM to analyze the problems that appear when one wants to redistribute an already existing input to a group of…
Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approa... more Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approach for simulating stochastic transient networks of bulk queues that often arise in consolidation networks. A set of approximations was proposed in order to produce a computationally tractable algorithm. This paper describes the numerical methods actually used for calculating probabilities and the results of extensive numerical experiments which test the accuracy of those approximations and the overall efficiency of the procedure, vis-à ...
W e consider a network revenue management problem where customers choose among open fare products... more W e consider a network revenue management problem where customers choose among open fare products according to some prespecified choice model. Starting with a Markov decision process (MDP) formulation, we approximate the value function with an affine function of the state vector. We show that the resulting problem provides a tighter bound for the MDP value than the choice-based linear program. We develop a column generation algorithm to solve the problem for a multinomial logit choice model with disjoint consideration sets (MNLD). We also derive a bound as a by-product of a decomposition heuristic. Our numerical study shows the policies from our solution approach can significantly outperform heuristics from the choice-based linear program.
Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived fr... more Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived from an application in less-than-truckload trucking, as a dynamic resource allocation problem. Drivers are aggregated into groups characterized by an attribute vector which capture the important attributes required to incorporate the work rules. The problem is very large: over 5,000 drivers and 30,000 loads in a four-day planning horizon. We formulate a problem that we call the heterogeneous resource allocation problem, which is more ...
Transportation Research Part E: Logistics and Transportation Review, 2008
The drayage services between a container terminal and the origin (or destination) of a shipment a... more The drayage services between a container terminal and the origin (or destination) of a shipment account for a significant portion of the total transportation cost. They are the key sources of shipment delays, road congestions, and disruptions in the international logistics network. Such a situation is even worse when the drayage services involve cross-border issues. Using Hong Kong, the busiest port in the world, as an example, we illustrate the challenges and issues in managing drayage activities in hub cities. We show that managing cross-border drayage container transportation is a very challenging problem because not only individual resources (e.g., driver, tractor, and chassis) but also the composites of them (e.g., the driver-tractor-chassis triplets) need to be managed simultaneously. The problem is further complicated by the regulatory * Corresponding author. † The authors would like to thank the Research Grants Council of Hong Kong to support the research through Grant 612206. policies which govern the cross-border activities. We use an attribute-decision model for this problem and implement an adaptive labeling algorithm to solve it. We conduct numerical experiments to evaluate the system performances under various regulatory policies. The results show that the benefit gained by relaxing the regulatory policies is significant.
Transportation Research Part B: Methodological, 1986
Relatively simple iterative procedures are developed for simulating the queue length distribution... more Relatively simple iterative procedures are developed for simulating the queue length distribution for transient bulk atrival, bulk service queues. The method allows the study of holding strategies where the length of time a vehicle is held can depend on both the length of the queue and how long the vehicle has been held. The system is modeled in discrete time, and a series of numerical experiments are presented that examine the errors introduced by this discretixation. *Financial support provided by CAPES and lnstituto Tecnol6gico de Aeronautica, Brazil.
Abstract A numerical approach is presented for determining the waiting time distribution in a tra... more Abstract A numerical approach is presented for determining the waiting time distribution in a transient bulk-arrival, bulk-service queue. Vehicle departures from the queue are governed by a general dispatch strategy that includes holding with a variable release function and vehicle cancellations. The waiting time distribution of a customer (in a group) arriving at a given point in time is calculated by simulating the process in discrete time and determining at each step the probability the customer has left the system. The dispatch strategies ...
Optimization models are sometimes promoted because they provide "optimal" solutions as defined by... more Optimization models are sometimes promoted because they provide "optimal" solutions as defined by a cost model. Simulation models, by contrast, are guided by rules that are specified by experts in operations. While these may seem heuristic in nature, they often reflect issues that are difficult to capture in a cost-based objective function. "Optimizing simulators" combine the intelligence of optimization with the flexibility of simulation in the handling of system dynamics, but still suffer from the limitation that the behavior is entirely determined by a cost model. In this paper, we show how a cost-based model can be guided through a set of low-dimensional patterns which are essentially simple rules determined by a domain expert. Patterns are incorporated through a penalty term, scaled by a coefficient that controls that tradeoff between minimizing costs and minimizing the difference between model behavior and the exogenous patterns.
Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approa... more Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approach for simulating stochastic transient networks of bulk queues that often arise in consolidation networks. A set of approximations was proposed in order to produce a computationally tractable algorithm. This paper describes the numerical methods actually used for calculating probabilities and the results of extensive numerical experiments which test the accuracy of those approximations and the overall efficiency of the procedure, vis-à ...
ABSTRACT Abstract We address the problem of combining a cost-based simulation model, which makes ... more ABSTRACT Abstract We address the problem of combining a cost-based simulation model, which makes decisions over time by minimizing a cost model, and rule-based policies, where a knowledgeable user would like certain types of decisions to happen with a specified frequency when averaged over the entire simulation. These rules are designed to capture issues that are dicult,to quantify as costs, but which produce more realistic behaviors in the judgment of a knowledgeable user. We consider patterns that are specified as averages over time, which have to be enforced in a model that makes decisions while stepping through time (for example, while optimizing the assignment of resources to tasks). We show how an existing simulation, as long as it uses a cost-based optimization model while stepping through time, can be modified to more closely match exogenously specified patterns.
In this paper we develop a new method for solving Dynamic Resource Allocation Problems DRAP that ... more In this paper we develop a new method for solving Dynamic Resource Allocation Problems DRAP that occur in the operation of freight transportation systems. Such problems involve the allocation of resources to perform tasks over a discrete-time, dynamic network. Our particular interest is in ultra-large scale problems, that involve managing thousands of resources such as drivers or vehicles. We focus on problems with dynamic attributes, where the characteristics of the resource may change as it handles each task. We provide a general and very exible formulation, and provide a solution based on the principle of dynamic programming. A newly proposed linearization approximation is shown to provide high quality solutions with reasonable execution times. The technique is illustrated in the context of the driver management problem for a large motor carrier.
Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived fr... more Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived from an application in less-than-truckload trucking, as a dynamic resource allocation problem. Drivers are aggregated into groups characterized by an attribute vector which capture the important attributes required to incorporate the work rules. The problem is very large: over 5,000 drivers and 30,000 loads in a four-day planning horizon. We formulate a problem that we call the heterogeneous resource allocation problem, which is more ...
Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approa... more Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approach for simulating stochastic transient networks of bulk queues that often arise in consolidation networks. A set of approximations was proposed in order to produce a computationally tractable algorithm. This paper describes the numerical methods actually used for calculating probabilities and the results of extensive numerical experiments which test the accuracy of those approximations and the overall efficiency of the procedure, vis-à ...
Proceedings. 2005 IEEE International Joint Conference on Neural Networks, 2005., 2005
The allocation of human and physical resources over time is a fundamental problem that is central... more The allocation of human and physical resources over time is a fundamental problem that is central to management science. In this paper we review a mathematical model of dynamic resource allocation that is motivated by problems in transportation and logistics. In principle, problems of this type can be solved via dynamic programming. However, three "curses of dimensionality" give rise to intractable computational requirements. We present computationally efficient approximate dynamic programming algorithms developed by the first author and collaborators for application to problems in freight transportation. We discuss how these algorithms address the three curses of dimensionality and how they relate to other independent threads of research on mathematical programming and approximate dynamic programming.
EURO Journal on Transportation and Logistics, 2012
Deterministic optimization has enjoyed a rich place in transportation and logistics, where it rep... more Deterministic optimization has enjoyed a rich place in transportation and logistics, where it represents a mature field with established modeling and algorithmic strategies. By contrast, sequential stochastic optimization models (dynamic programs) have been plagued by the lack of a common modeling framework, and by algorithmic strategies that just do not seem to scale to real-world problems in transportation. This paper is designed as a tutorial of the modeling and algorithmic framework of approximate dynamic programming, however our perspective on approximate dynamic programming is relatively new, and the approach is new to the transportation research community. We present a simple yet precise modeling framework that makes it possible to integrate most algorithmic strategies into four fundamental classes of policies, the design of which represent approximate solutions to these dynamic programs. The paper then uses problems in transportation and logistics to indicate settings in which each of the four classes of policies represent a natural solution strategy, highlighting the fact that the design of effective policies for these complex problems will remain an exciting area of research for many years. Along the way, we provide a link between dynamic programming, stochastic programming and stochastic search.
2013 46th Hawaii International Conference on System Sciences, 2013
ABSTRACT Given the importance of handling high levels of uncertainty from renewables in the unit ... more ABSTRACT Given the importance of handling high levels of uncertainty from renewables in the unit commitment problem, there has been increased attention given to the use of stochastic programming methods. Since these are computationally very demanding, there is a need for new approximations. We propose to use a point forecast of energy from wind and loads, where the point forecast is chosen and adapted by simulation to produce a robust solution. Traditionally, point forecasts represent an expectation. In our work, we suggest that we can use an appropriately chosen quantile of the forecast distribution which is optimized within a stochastic environment. The result is a policy search algorithm built around a point forecast, which is easily implement able using standard industry models and algorithms.
Proceedings of the 2010 Winter Simulation Conference, 2010
We describe an adaptation of the knowledge gradient, originally developed for discrete ranking an... more We describe an adaptation of the knowledge gradient, originally developed for discrete ranking and selection problems, to the problem of calibrating continuous parameters for the purpose of tuning a simulator. The knowledge gradient for continuous parameters uses a continuous approximation of the expected value of a single measurement to guide the choice of where to collect information next. We show how to find the parameter setting that maximizes the expected value of a measurement by optimizing a continuous but nonconcave surface. We compare the method to sequential kriging for a series of test surfaces, and then demonstrate its performance in the calibration of an expensive industrial simulator.
The Spherical Frontier DEA Model (SFM) (Avellar et al., 2007) was developed to be used when one w... more The Spherical Frontier DEA Model (SFM) (Avellar et al., 2007) was developed to be used when one wants to fairly distribute a new and fixed input to a group of Decision Making Units (DMU's). SFM's basic idea is to distribute this new and fixed input in such a way that every DMU will be placed on an efficiency frontier with a spherical shape. We use SFM to analyze the problems that appear when one wants to redistribute an already existing input to a group of…
Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approa... more Abstract In a companion study, Simão and Powell have introduced a numerical, discrete-time approach for simulating stochastic transient networks of bulk queues that often arise in consolidation networks. A set of approximations was proposed in order to produce a computationally tractable algorithm. This paper describes the numerical methods actually used for calculating probabilities and the results of extensive numerical experiments which test the accuracy of those approximations and the overall efficiency of the procedure, vis-à ...
W e consider a network revenue management problem where customers choose among open fare products... more W e consider a network revenue management problem where customers choose among open fare products according to some prespecified choice model. Starting with a Markov decision process (MDP) formulation, we approximate the value function with an affine function of the state vector. We show that the resulting problem provides a tighter bound for the MDP value than the choice-based linear program. We develop a column generation algorithm to solve the problem for a multinomial logit choice model with disjoint consideration sets (MNLD). We also derive a bound as a by-product of a decomposition heuristic. Our numerical study shows the policies from our solution approach can significantly outperform heuristics from the choice-based linear program.
Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived fr... more Abstract We consider an aggregated version of a large-scale driver scheduling problem, derived from an application in less-than-truckload trucking, as a dynamic resource allocation problem. Drivers are aggregated into groups characterized by an attribute vector which capture the important attributes required to incorporate the work rules. The problem is very large: over 5,000 drivers and 30,000 loads in a four-day planning horizon. We formulate a problem that we call the heterogeneous resource allocation problem, which is more ...
Transportation Research Part E: Logistics and Transportation Review, 2008
The drayage services between a container terminal and the origin (or destination) of a shipment a... more The drayage services between a container terminal and the origin (or destination) of a shipment account for a significant portion of the total transportation cost. They are the key sources of shipment delays, road congestions, and disruptions in the international logistics network. Such a situation is even worse when the drayage services involve cross-border issues. Using Hong Kong, the busiest port in the world, as an example, we illustrate the challenges and issues in managing drayage activities in hub cities. We show that managing cross-border drayage container transportation is a very challenging problem because not only individual resources (e.g., driver, tractor, and chassis) but also the composites of them (e.g., the driver-tractor-chassis triplets) need to be managed simultaneously. The problem is further complicated by the regulatory * Corresponding author. † The authors would like to thank the Research Grants Council of Hong Kong to support the research through Grant 612206. policies which govern the cross-border activities. We use an attribute-decision model for this problem and implement an adaptive labeling algorithm to solve it. We conduct numerical experiments to evaluate the system performances under various regulatory policies. The results show that the benefit gained by relaxing the regulatory policies is significant.
Transportation Research Part B: Methodological, 1986
Relatively simple iterative procedures are developed for simulating the queue length distribution... more Relatively simple iterative procedures are developed for simulating the queue length distribution for transient bulk atrival, bulk service queues. The method allows the study of holding strategies where the length of time a vehicle is held can depend on both the length of the queue and how long the vehicle has been held. The system is modeled in discrete time, and a series of numerical experiments are presented that examine the errors introduced by this discretixation. *Financial support provided by CAPES and lnstituto Tecnol6gico de Aeronautica, Brazil.
Abstract A numerical approach is presented for determining the waiting time distribution in a tra... more Abstract A numerical approach is presented for determining the waiting time distribution in a transient bulk-arrival, bulk-service queue. Vehicle departures from the queue are governed by a general dispatch strategy that includes holding with a variable release function and vehicle cancellations. The waiting time distribution of a customer (in a group) arriving at a given point in time is calculated by simulating the process in discrete time and determining at each step the probability the customer has left the system. The dispatch strategies ...
Optimization models are sometimes promoted because they provide "optimal" solutions as defined by... more Optimization models are sometimes promoted because they provide "optimal" solutions as defined by a cost model. Simulation models, by contrast, are guided by rules that are specified by experts in operations. While these may seem heuristic in nature, they often reflect issues that are difficult to capture in a cost-based objective function. "Optimizing simulators" combine the intelligence of optimization with the flexibility of simulation in the handling of system dynamics, but still suffer from the limitation that the behavior is entirely determined by a cost model. In this paper, we show how a cost-based model can be guided through a set of low-dimensional patterns which are essentially simple rules determined by a domain expert. Patterns are incorporated through a penalty term, scaled by a coefficient that controls that tradeoff between minimizing costs and minimizing the difference between model behavior and the exogenous patterns.
Uploads
Papers by Hugo Simao