Airline Industry Linear Programming Models Tourism Essay

Introduction

Operational Research, also known as Management Science, has been considered to be a subject that has a profound influence on the airline industry (Yu and Yang, 1988, an overview of issues). Nowadays, major airlines have their own OR departments that cooperates with other department to improve the company’s operation while small airlines employ OR personnel to provide professional consulting. (an over view of issue).

Due to the competition from peer air carriers and ground transportation companies as well, the airline market becomes increasingly fierce. How to optimize airline’s profit, and thus gain and keep leading edge in the airline industry is the main issue faced by the managers of every airline company ( or in airline industry). Generally, this goal (maximizing the profit) can be achieved by two means: the first way is reducing the operating cost by optimizing the daily operations; the other way is maximizing the revenue by balancing the number of customers and the flight price. These two sub-objectives are the main concerns of Operational Research in the airline industry.

Specifically, according to previous studies, Operational Research practices in Airlines cover a rang of domains: scheduling planning~~~~~~. As an Operational Research analyst working in airline industry, ones are expected to have and use the following techniques and knowledge in their daily work: mathematical programming, simulation, statistic analysis, problem structuring methods and ~~~~~. This essay will describe the OR techniques in details, followed by case studies to illustrate how they can be applied in practice.

First case study

OR techniques-linear programming

Linear programming models are one of the most widely used models in OR ( Paul Williams,1999). In a survey of fortune 500 firms, 85% of the respondents said they had used linear programming (Wayne Winston, 2004). Application in airline.

Linear programming in OR has one essential feature that it involves a set of mathematical relationships (such as equations, inequalities, etc.) that represent more down-to-earth relationships in the real world (Paul Williams, 1999). Specifically, a linear programming model usually has three modules: firstly, it involves optimization. Models are built to maximize something or minimize something. Secondly, it contains constraints that correspond to real relationships between variables. The last one is the input data such as costs, resource availabilities, technological coefficients, etc.

Regression analysis

In today’s industrial processes, there is no shortage of “information” (Norman R.Draper and Harry Smith, 1998). Adequate information provides needed supports for examining the effects that some variables exert on others. Regression analysis is a technique that can be used to identify such effects by means of mathematical functions.

Case

Background and problem description

This section will explore how linear programming together with regression analysis can be applied in the airline industry by illustrating a real case.

In airline industry, fuel cost is regard as one of the major operating cost, which can account up to 20% of total operating cost. It is nature for every airline company to hammer at reducing its fuel cost. Middle East Airlines (MEA) is no exception. The particular fuel cost problem addressed here is to minimizing the total fuel cost by taking different prices at different airport into consideration.

The original fuel uplifting policy at MEA was that the pilots determine where and how much to uplift fuel by its own experience. Due to the absence of any formal decision making tool for minimizing the fuel cost, the pilots usually only considered a single flight leg to decide where to uplift. However, this strategy was myopic because it ignoreed the fact that the fuel prices varied from airport to airport.

Before building LP model, it was assumed that the aircraft can uplift enough fuel for more than one single flight leg according to its capacity, which consists with the facts. It might seem to be a good method to simply uplift as much fuel as one aircraft can carry at the airport with lowest price. However, this was myopic as the original policy because more fuel an aircraft carried that increased the aircraft’s take-off weight could result in more fuel consumed over the flight. Thus, a LP model was introduced to balance the trade-off between the savings of fuel prices and the cost of higher take-off weight.

OR application

Before building model, regression analysis was used to identify the relationship between the take-off weight and the fuel consumed over the flight by based on the historical data. An approximately linear relationship was found and would be used in the LP model.

Then, based on the data provided by EMA, a linear programming model was established to determine the amount of fuel to be uplifted by the aircraft at each airport by minimizing the total fuel cost and by being subject to the constraints on take-off weight, landing weight, aircraft’s fuel capacity and safety fuel at landing. After that, LINGO, a software solver for Linear Programming, was applied to solve the model.

By adopting this new fuel uplifting strategy, MEA had achieved amount to $87,923 per week savings in the fuel cost on average, which was 10% of overall fuel cost.

Lesson learned

This case indicates that LP is widespread used not only because it is effective and powerful in handling operational problems but also because it is comparably time saving and labor saving. For its frequently usages, LP is regarded as a basic technique a OR analyst is expected to master.

two OR techniques-LP model and Regression analysis can be applied into airline industry effectively.

Second case

Simulation

‘The process of designing a model of a concrete system and conducting experiments with this model in order to understand the behaviour of concrete system and/or to evaluate various strategies for the operation of the system.’ (Shannon, 1975)

Statistic

Background and problem description

In the following section, a real case that simulation tool and statistic analysis are used to solve an airline passenger congestion problem will be analyzed.

Customers are accustomed to using air transportation as a primary means of transportation. (using fuzzy cognitive). The number of airline passengers keeps rapid growth, while airport capacity is growing slowly. Under this circumstance, the Airline passenger terminal congestion is becoming a crucial problem faced by top management of every airline.

In order to avoid unsatisfactory levels of customer service caused by congestion, Air Canada, IBM Corporation and Systems Modeling Corporation design and develop simulation models to plan for advancement in existing or proposed terminal facilities. A simulation model of passenger and baggage processing at Toronto airport was developed to examine the impact of introducing new information technology capabilities.

The simulation model for Toronto airport is developed based on IBM journey management library (IBM JML) that is a custom designed modeling template developed with the Arena simulation software for simulating airline passengers pressing at an airport. IBM JML contains a set of reusable modules-each of them simulates a single component of passenger processing at an airport. Technically, there are three types of modules. First, data modules function to define the basic information about passenger and flights (e.g. passenger arrival pattern and flight schedule). Second, logic modules function to simulate passenger flow (e.g. queuing area where passengers wait and service location where passengers are serviced. The last type is the group of process modules, which are developed to define the service process (e.g. issue ticket, issue boarding pass). And a service process normally has a service time following a particular distribution.

(basic computer programming knowledge is needed)

OR application

The model aimed to simulate airport passenger processing at terminal two and baggage handling operations for Toronto’s Lester B.Pearson Airport, and then to examine the impact of introducing new information technology capabilities.

In the simulation, passengers and baggages were identified as the key entities that go through the airport passenger service process.

As the first step, simulation model of Toronto airport process was constructed by selecting modules that represent airport components and service processes. Toronto airport provides input data for the simulation model from current observations and historical patterns. Then, the actual airline performance is used to validate the simulation model. Here, some statistic analysis knowledge (i.e. Chi-square and Kolmogorov-Smirnov tests) is applied to test how close the actual data from observation is to the simulation results. As a result, the simulation model is quite satisfactory with most statistics of simulation model and actual performance differing by less than 5%. Finally, the model is used to simulate the airport operations by take new information technology capabilities into consideration. Impacts of introducing new capabilities are gained.

Lesions learned

This model suggests that simulation is a ~~~technique that an OR analyst can applied when he or she works in the airline industry. Simulation is a effective method to address processing optimization problems. In applying this method, it is worthy to emphasize (zai san)that the simulation model must be represent the real problem as accurately as possible and that a ~~validation is always needed.

Third case

Describing Techniques

Three OR techniques will be described first in this section.

PSMs

In a networked society, a big organization such as an airline company has to face some ill-constructed problem situations that are so socially complex that it is not even easy to identify the objectives. Under this circumstance, hard OR including mathematical programming, simulation and etc. itself cannot solve the problem alone. PSMs can contribute to solve such complexities.(hamer and champy,1993)

Problem structuring methods use models with little or no quantification to help group decision making (RA). PSMs contain methods such as strategic options development and analysis (SODA), soft systems methodology (SSM), strategic choice approach (SCA), robustness analysis, and drama theory. (RA!!!!@!@#) The methods are various, but the problems addressed by PSMs share same characteristics: technological complexity, human/social complexity, and divergence of values and interests (Jackson and keys,@@@@@). And the common soft OR principles (den Hengest and de Vreede, 2004) behind the PSMs can be summarized as follows:

Outcome acceptance: the results should be trusted by stakeholders.

Convergence among stakeholders: stakeholders should achieve a greate degree of shared understanding (Eden and Ackerman, 1998).

Understanding through models: the models should be easy to understand for non-modelling experts ( Hlupic and de Vreede, 2005).

Encouraging involvement: stakeholders’ involvement is the root of PSMs.

Model quality: models should be as accurate as possible.

Efficiency: the time a PSM take should be reasonable.

In practice, a combination of several PSMs rather than one single PSM is needed to handle a complex situation in many cases. And sometimes, it is even hard to identify which PSM is applied. However, den Hengst, de Vreede and R Maghnoji (2007) believe that a method satisfied all the six soft OR principles can be regarded as an effective PSM. Collaborative simulation that has been applied in the following case is considered to be a PSM by den Hengst, de Vreede and R Maghnoji (2007)

Collaborative simulation

Collaborative simulation is a method introduced by den Hengst, de Vreede and R Maghnoji (2007) to refer to the combinations of the DES and GSS. GSSs provide tools and techniques to help with structuring of activities and generation of ideas ( Tyran et al, 1992)¼Œ which has three main characteristics-anonymity, parallel input and group memory ¼ˆFjermestad andHoltz¼Œ 1998-1999).

Either DES or GSS itself does not embed all soft OR principles, but the combination of them (i.e. collaborative simulation) seems to embed all the principles¼ˆPlease refer to appendix for details¼‰. Thus, collaborative simulation is considered to be a effective PSM (den Hengst, de Vreede and R Maghnoji, 2007).

a case will be analyzed to show how soft OR principles will be applied to help with constructing a simulation model.

Case study

Background and description of problem

This case is about a large airline carrier in Dutch which carries more than 15 million passengers and 600,000 tons of cargo. At that time, the airline faced a complex situation: the top management had a debate on whether the existing hub was big enough to handle increasing cargo flows; meanwhile, the airline would have to move its warehouses to another location at the airport and reconstruct them as the home airport wanted to expand its terrain for passenger handing at the hub.

The manage team thought that this situation was an opportunity to redesign the cargo-handling processes. Therefore, a simulation project was conducted to explore the future optimal cargo-handling processes.

As a result of several debates, the problem was defined as follows: ‘the management team lacks shared insight into possibilities and limitations of the cargo-handling processes, both at current and alternative locations, given growing cargo flows and expansion plans of the home airport.’ Complex cargo-handling processes (representing technologically complexity) and the management team lack shared insight (reflecting social complexity and divergence of values)-three characteristics of problems addressed by PSMs-strongly indicated that PSMs can and need to be applied in this case.

OR application

The airline applied collaborative simulation in several steps:

Firstly, operational managers and staff were interviewed to identify simulation initial objectives, output variables and model structure. Based on this information, an empirical simulation model was developed by using AutoMod-a software package for simulation. The reason of choosing AutoMod is that it has the strong ability to quickly build a true-to-scale animation model (####), which made it easier for managers to get acquainted with simulation model. That obeyed the soft OR principle-understanding through model.

Secondly, two validation sessions with operational managers was hold. Most of the managers lacked real operational insight, thus felt the model results should be validate with operational staff. The staff confirmed that the simulation results represented the reality. With this endorsement, the operational managers finally validated the simulation model.

Thirdly, a management team session is hold. In this session, management ream used GroupSystems’ brainstorming modules to identify more than 10 alternatives. During verbal discussions, management team established shared understanding that 3modt preferred alternative were selected for further analysis.

Fourthly, second management team session is hold to explore alternatives. OR consultants constructed three models of the selected alternatives and analyze the output data. According to these outcomes of the models, management team used brainstorming modules to discuss the perceived advantages and disadvantages of each alternative.

Finally, a decision was made about choosing a direction for the company in the final management team session.

Lessons learned

Applying PSMs is a challenging task. To begin with, the role of OR consultant is more than a “modeler” or “analyst”(RA) when apply PSMs. In this case, we can see that OR consultants play the role of observer and participant: as an observer, OR consultant is require to observe group session and collect valuable information (both quantitative and qualitative); as a participant, one is responsible for scheduling and facilitating group sessions. In addition, when facing a wicked problem, a single PSM is seldom the answer. Such problems need a combination of several PSMs and other hard OR methods, which require a OR consultant not only maters the PSM approaches but also comprehends the purpose and the soft OR principles behind them.

Conclusion

According to the analysis above, we may reach the conclusion safely that as a qualified OR analyst, one must acknowledge both hard OR techniques including traditional methods such as mathematical programming, simulation, etc. and soft OR techniques such as PSMs. It worth emphasize time and again that it is more important to master the root principles and purpose behind a method and have the ability to combine several methods to handling wicked situation as the problems faced by OR analyst are increasingly technological and social complex that any single method is seldom the answer.