Huei Chuen Huang

Short-term booking of air cargo space

This paper proposes a stochastic dynamic programming model for a short-term capacity planning model for air cargo space. Long-term cargo space is usually acquired by freight forwarders or shippers many months ahead on a contract basis, and usually the forecasted demand is unreliable. A re-planning of cargo space is needed when the date draws nearer to the flight departure time. Hence, for a given amount of long-term contract space, the decision for each stage is the quantity of additional space required for the next stage and the decision planning model evaluates the optimal cost policy based on the economic trade-off between the cost of backlogged shipment and the cost of acquiring additional cargo space. Under certain conditions, we show that the return function is convex with respect to the additional space acquired for a given state and the optimal expected cost for the remaining stages is an increasing convex function with respect to the state variables. These two properties can be carried backward recursively and therefore the optimal cost policy can be determined efficiently.

A successive convex approximation method for multistage workforce capacity planning problem with turnover

Workforce capacity planning in human resource management is a critical and essential component of the services supply chain management. In this paper, we consider the planning problem of transferring, hiring, or firing employees among different departments or branches of an organization under an environment of uncertain workforce demands and turnover, with the objective of minimizing the expected cost over a finite planning horizon. We model the problem as a multistage stochastic program and propose a successive convex approximation method which solves the problem in stages and iteratively. An advantage of the method is that it can handle problems of large size where normally solving the problems by equivalent deterministic linear programs is considered to be computationally infeasible. Numerical experiments indicate that solutions obtained by the proposed method have expected costs near optimal.

A Markov model for single-leg air cargo revenue management under a bid-price policy

In this paper, we consider the capacity allocation problem in single-leg air cargo revenue management. We assume that each cargo booking request is endowed with a random weight, volume and profit rate and propose a Markovian model for the booking request/acceptance/rejection process. The decision on whether to accept the booking request or to reserve the capacity for future bookings follows a bid-price control policy. In particular, a cargo will be accepted only when the revenue from accepting it exceeds the opportunity cost, which is calculated based on bid prices. Optimal solutions are derived by maximizing a reward function of a Markov chain. Numerical comparisons between the proposed approach and two existing static single-leg air cargo capacity allocation policies are presented.

SimMan-A simulation model for workforce capacity planning

Motivated by the difficulty in predicting the actual performance of workforce capacity planning solutions obtained from mathematical models, we develop a simulator, SimMan, to serve as a test bed for evaluating the effectiveness and robustness of different planning options and assignment rules. To facilitate the ease of use, the simulator is developed in C++ language with a modular structure so that different users or planners can customize the modules according to their specifications. We further develop a mathematical model for the workforce capacity planning problem, where the uncertainty of the demand is handled by the safety stock concepts. We also suggest a number of practical planning alternatives and assignment rules based on information from IBM. These rules are implemented as modules in SimMan and tested numerically to provide managerial insights.

Development of timed colour Petri net simulation models for air cargo terminal operations

This paper presents the development and the application of simulation models for air cargo terminal operations. As air cargo volume is fast increasing in recent years, air cargo terminals which handle cargos for carriers face challenges to streamline their operations. Air cargo terminals employ diverse storage media and are also equipped with various material handling equipments. Due to the complex and stochastic nature of terminal operations, which makes formulation of analytical model difficult, in this study, simulation models are developed to analyze the air cargo terminal operations. We first employ timed CPN (Colour Petri nets) to model the terminal operations, and then the corresponding simulation model will be developed. In order to validate this simulation model, we run the model based on actual cargo retrieval schedules obtained from an air cargo terminal. The results show that this model provides a good estimate of the terminal performance. The model is then employed to analyze the performance of the airline assignment policy for the objective of minimizing the cargo processing times, and the results show that the proposed policy significantly improves the operational performance of air cargo terminals. Second, the simulation model is employed to analyze the performance of an AS/ RS (Automated Storage and Retrieval System), and the performance of the proposed storage policy is compared with that of the current policy. The simulation experiment also validates the optimal parameter value from the analytic model. The developed models simulate air cargo terminal operations effectively and efficiently, and will aid future studies in the design and control of terminal operations.

Discrete event simulation model for airline operations: SIMAIR

SIMAIR is a C++ based research tool meant for the simulation of airline operations. It provides a means for devising and evaluating various airline recovery mechanisms to handle disruptions, and can also be used as a tool to evaluate the performance of a given schedule of operations. The performance of a given recovery mechanism is quantified for research and evaluation purposes.

Asymptotic expected performance of some TSP heuristics: An empirical evaluation

Abstract This paper presents an empirical study on the performance of some heuristics for the travelling salesman problem. Our study indicates that the expected tour lengths through n points in a unit square produced by these heuristics are all proportional to SQRT(n) asymptotically. In this study, the average tour lengths generated by each heuristic are approximated by a linear function of SQRT(n). The model parameters are estimated by the linear regression method. The possible effect of different random number generators, which are used to perform the statistical analysis, is also investigated.

A two-echelon inventory system with transportation capacity constraint

Abstract Consider a one-warehouse multi-retailer system under constant and deterministic demand, which is subjected to transportation capacity for every delivery period. To search for the best stationary zero inventory ordering (ZIO) policy, or the best power-of-two policy, or the best nested policy, the problem is formulated as a 0–1 integer linear program in which the objective function comprises of a fixed transportation cost whenever a delivery is made and the inventory costs for both the warehouse and retailers. To overcome the transportation capacity limitation, we extend the policies to allow for staggering deliveries. It is shown that with transportation capacity constraint the non-staggering policy can have its effectiveness close to 0% from the best staggering policy and the power-of-two policy with staggering allowed can have its effectiveness close to 0% from the optimal policy. Nevertheless in general, the power-of-two policy fairs well on a number of randomly generated problems. To solve the large distribution network problem, an efficient heuristic based on the power-of-two policy with staggering of deliveries is suggested.

The economic lot scheduling problem under extended basic period and power-of-two policy

The economic lot scheduling problem schedules the production of several different products on a single machine over an infinite planning horizon. In this paper, a nonlinear integer programming model is used to determine the optimal solution under the extended basic period and power-of-two policy. A small-step search algorithm is presented to find a solution which approaches optimal when the step size approaches zero, where a divide-and-conquer procedure is introduced to speed up the search. Further a faster heuristic algorithm is proposed which finds the same solutions in almost all the randomly generated sample cases.

A k-product uncapacitated facility location problem

Abstract A k-product uncapacitated facility location problem can be described as follows. There is a set of demand points where clients are located and a set of potential sites where facilities of unlimited capacities can be set up. There are k different kinds of products. Each client needs to be supplied with k kinds of products by a set of k different facilities and each facility can be set up to supply only a distinct product with a non-negative fixed cost determined by the product it intends to supply. There is a non-negative cost of shipping goods between each pair of locations. These costs are assumed to be symmetric and satisfy the triangle inequality. The problem is to select a set of facilities to be set up and their designated products and to find an assignment for each client to a set of k facilities so that the sum of the setup costs and the shipping costs is minimized. In this paper, an approximation algorithm within a factor of 2 k + 1 of the optimum cost is presented. Assuming that fixed setup costs are zero, we give a 2 k - 1 approximation algorithm for the problem. In addition we show that for the case k = 2 , the problem is NP-complete when the cost structure is general and there is a 2-approximation algorithm when the costs are symmetric and satisfy the triangle inequality. The algorithm is shown to produce an optimal solution if the 2-product uncapacitated facility location problem with no fixed costs happens to fall on a tree graph.