Disasters such as hurricanes, earthquakes and floods continue to have devastating socioeconomic impacts and endanger millions of lives. Shelters are safe zones that protect victims from possible damage, and evacuation routes are the paths from disaster zones toward shelter areas. To enable the timely evacuation of disaster zones, decisions regarding shelter location and routing assignment (i.e., traffic assignment) should be considered simultaneously. In this work, we propose a risk-averse stochastic programming model with a chance constraint that takes into account the uncertainty in the demand of disaster sites while minimizing the total evacuation time. The total evacuation time reflects the efficacy of emergency management from a system optimal (SO) perspective. A conditional value-at-risk (CVaR) is incorporated into the objective function to account for risk measures in the presence of uncertain post-disaster demand. We resolve the non-linear travel time function of traffic flow by employing a second-order cone programming (SOCP) approach and linearizing the non-linear chance constraints into a new mixed-integer linear programming (MILP) reformulation so that the problem can be directly solved by state-of-the-art optimization solvers. We illustrate the application of our model using two case studies. The first case study is used to demonstrate the difference between a risk-neutral model and our proposed model. An extensive computational study provides practical insight into the proposed modeling approach using another case study concerning the Black Saturday bushfire in Australia.

The air transport industry is an important branch of the intelligent transportation system (ITS). It is widely admitted that modern ITS technologies and advanced management methods, such as fleet assignment, aircraft maintenance routing, and crew scheduling, can significantly increase an airline's market share and profit, and also improve customer satisfaction. This paper studies a new airline stochastic fleet assignment problem with random passenger demands under risk aversion. The objective is to maximize the expected total profit at a certain level of risk avoidance (i.e., conditional value-at-risk). To solve this problem, we present a risk-averse two-stage stochastic mixed-integer programming model. The first stage mainly deals with tactic level decisions: assigning aircraft families (e.g., Airbus A380 family) to flight legs. The operational level decisions are made in the second stage to efficiently assign aircraft types (e.g., Airbus A380-800 or A380-800F) to flight legs while meeting the family assignment plan developed in the first stage. Then, a sample average approximation algorithm is proposed to solve the stochastic programming problem considering risk aversion. A realistic international airline's numerical experiment is conducted to illustrate the efficiency of the proposed two-stage stochastic programming model and algorithm.

Air traffic flow management is one of the important operations in terminal airports heavily relying on advanced intelligence transportation techniques. This work considers a two-stage runway scheduling problem given a set of flights with uncertain arrival times. The first-stage problem is to identity a sequence of aircraft weight classes (e.g., Heavy, Large and Small) that minimizes runway occupying time (i.e., makespan). Then the second-stage decision is dedicated to scheduling the flights as punctually as possible after their arrival times realized, which translates into determining a sequence of flights for each aircraft category such that the total deviation time imposed on the flights is minimized. Instead of an exactly known probability distribution, information on uncertain parameters is limited (i.e., ambiguous), such as means, mean absolute deviations and support set of random parameters derived from historical data. Under this information on the random parameters, an ambiguous mixed-integer stochastic optimization model is proposed. For such a problem, we approximately construct a worst-case discrete probability distribution with three possible realizations per random parameter, and adopt a hybrid sample average approximation algorithm. To illustrate the effectiveness and efficiency of the proposed model and algorithm, extensive numerical experiments are carried out.

Assembly line worker assignment and balancing problem (ALWABP) is committed to assign assembly tasks and workers to work stations that constitute the assembly line. Contrary to the hypothesis that, for each worker, the processing time of task is deterministic, we consider an uncertain version for the ALWABP in which the processing time is a stochastic parameter and exact probability distribution is known. We incorporate the risk measure (e.g., Conditional Value-at-Risk (CVaR)) into the objective function rather than the risk-neutral one based on expectation. The objective of our risk-averse ALWABP is to find an assignment of tasks to workers (to stations) such that the maximum cycle time (i.e., the maximum completion time among all workers) and the value of risk measure are minimum. For the handling problem is of large scale and taking an insight into the stochastic intrinsic nature, we tentatively propose a sample average approximate (SAA) algorithm to approximately solve the ALWABP with risk aversion.

This work addresses the problem of assigning airline check-in employees to tasks related to departing flights under uncertain circumstance at an international terminal of a large airport. The uncertainty of flight departing time mainly stems from delays and traffic control. Airlines specifics the minimum and target number of staff for each flight task in each time period. Taking uncertain tasks starting time into account and introducing risk measure (e.g., conditional value-at-risk) into optimization objective function, we present a risk-averse two-stage stochastic model with the mixed-integer recourse that minimizes both staffing costs and risk measurement to respond the time uncertainty. The first-stage decision is to make personal-area allocation plan; and the particular personnel-task scheduling decisions are determined in the second stage as the actual flight departing times are realized. To solve the NP-hard problem, a progressive hedging algorithm is proposed. Numerical experiments are carried out to demonstrate the efficiency and effectiveness of the proposed approach as compared with a sample average approximation method. Finally, we utilize a real-world case to illustrate how the proposed method can be used to obtain a near-optimal solution for practical problem.

This work studies an employee scheduling problem considering average and worst case simultaneously under uncertainty demand (i.e., customer traffic) in retailing. Stochastic employee scheduling comprises two stages, the here-and-now decision (i.e., first-stage), before the actual demand is known, is to allocate number of full-time employees to shifts by using some forecast or empirical data; the wait-and-see decision (i.e., second-stage) involving takes some recourse actions, such as recruits part-time employees and extends shift length of fulltime employees (i.e, overtime shift), since the actual demand realization. In order to avoid large customers loss arises as extremely worst scenario happens, the worst case cost is taken into consideration in the second-stage objective function. The objective, in our model, is minimizing the fixed cost of fulltime employee and weight sum of average and worst case cost of recourse action. Thus, the model offers some flexibility to decision makers by putting relative emphasis on the average and worst case’s recourse cost in term of different scheduling schemes. To solve this problem, a modified sample average approximation algorithm, which combining the clustering technique, is proposed.