Multi criteria analyses for managing motorway company facilities: The decision support system SINERGIE
Introduction
Advanced computing applications are changing the way engineers interact with computers and many examples can be found in infrastructure management; infrastructure management itself concerns the full life cycle of the infrastructure, including the planning, design, construction, commissioning, operation, maintenance, rehabilitation and decommissioning. Engineers in charge of large-scale infrastructure management are now skilled in the craft of using computational tools for tasks such as numerical analysis, drafting, detail design and other aspects of project planning [1]. A lot of soft computing solutions have been proposed in the recent years to assist them in any step of the life-cycle of an infrastructure from design to maintenance. Thus, the impact of information and communication technologies (ICT) is a kernel factor in developing our modes of organisation, if not of our societies. Regardless of how humans are involved in systems nowadays, the systems are so complex that increasingly intricate and inescapable dynamic information processing systems are bound to emerge. Decision support systems (DSS) constitute one of the most significant classes of ICT tools and have been the subject of thorough investigations. DSS may rely on statistical, data-mining, knowledge discovery techniques, but also on operational research, theories of uncertainty and multi criteria techniques. The next paragraph gives a short panel of some works relative to decision-making aids in the area of infrastructure management.
Saridakis et al. [2] survey the application of soft computing techniques in engineering design. Within this context, fuzzy logic, genetic algorithms and artificial neural networks, as well as their fusion are reviewed in order to examine the capability of soft computing methods and techniques to effectively address various hard-to-solve design tasks and issues. In [3], the authors explain that infrastructure managers rely on capabilities of computer-aided design (CAD) and geospatial information systems (GIS) for making decisions during the implementation of engineering tasks. Engineers in infrastructure management must gain knowledge and skills in both CAD and GIS to perform these tasks. Interoperability is at the heart of their work and is seen as a solution to overcome the problems associated with heterogeneous environments. The interoperability may occur at different levels and for different purposes. In a recent paper, Dehlin and Olofsson [4] discuss the motivation to innovate and to introduce new ICT tools and working methods into the construction industry. In view of this, they explain that a new project-oriented evaluation model is developed with the purpose of providing for a structure and a work routine to be used by a multidisciplinary project team to evaluate the implications of realizing ICT investments in construction projects. They underline that new ICT tools in construction have to support sharing of information and give guidelines to evaluate their value or impact on the profitability of the project. Although primarily aimed at establishing future benefits and costs, their model may be used for follow-ups. Kostoulas et al. [5] highlight the role of ICT in construction project control. Integrating promising information technologies such as radio frequency identification (RFID) technology, mobile devices-PDA and web portals can help improve the effectiveness and the convenience of the information flow in the construction systems to control the supply chain. More marginally, Kostoulas et al. [6] present a decentralized trust model to enhance reliable information dissemination in large-scale disasters relief operations involving civil engineers.
Decision support systems (DSS) constitute probably the most cognitive ICT tools. A DSS is defined as the combination of data, information, and computer based tools and services working within a structured framework to improve the process and outcome of decision-making [7]. Vanier Dana [7] mentions six characteristics of a DSS: (1) explicit design to solve ill-structured problems; (2) powerful and easy-to-use interface; (3) ability to flexibly combine analytical models with data; (4) ability to explore the solution space by building alternatives; (5) capability of supporting a variety of decision-making styles and (6) allow interactive and recursive problem-solving. In the special issue in ITCON, DSS in infrastructure management [7], Vanier introduces a series of papers that illustrate the role of DSS in infrastructure management. Computer based tools play a key role: 3D visualization tools to plan and lay out urban green spaces, geographic information systems (GIS) [8], ontologies and more generally semantic web to promote interoperability in large scale integrated projects [9], [10], knowledge based decision support system [11], [12], user-friendly man-machine interface [13], planning software [13], client-server decision support architecture for project management [14].
Many water utilities are faced with the problem of ageing pipe networks and the associated increasing costs. That is why lots of DSS have been proposed in this field since the sixties. [15] presents a decision support system called PARMS-PRIORITY, a software application to support decisions regarding pipeline renewal prioritisation. The modules described are based on key decision-making tasks, such as: risk calculation, failure prediction, costs assessment, data exploration and scenario evaluation. In the same way, a method for estimating water network rehabilitation needs is proposed in [16]. [17] also proposes a DSS for infrastructure maintenance to water supply systems. Finally, [18] proposes a DSS for rehabilitation planning and optimisation of the maintenance of underground pipe networks of water utilities: the DSS performs reliability based life predictions of the pipes and determines the consequences of maintenance and neglect over time in order to optimize a rehabilitation policy. The aim of a DSS may also be to satisfy the high transparency requirement for financial controls and the information demand for forward financial planning [19], [20], [21]. The techniques behind these works are most often data mining, neural networks [21] and signal processing, e.g. data compression, data streams, noise filtering, from one hand [22], multi attribute utility theory and operational research (optimization problems with discrete and/or continuous variables) from the other hand [12]. [23] stimulates interest within the civil engineering research community for developing the next generation of applied artificial neural networks. Data-mining, knowledge discovery and automated learning techniques are of use in many civil engineering activities. Multi criteria approaches also provide relevant tools in civil engineering. For example, [24] concerns design governed by multiple objective criteria, which are conflicting in the sense of competing for common resources to achieve variously different performance objectives (financial, functional, environmental, esthetical, etc.). A multi criteria decision making (MCDM) strategy is proposed that employs a trade-off-analysis technique to identify compromise designs for which the competing criteria are mutually satisfied in a Pareto-optimal sense.
Decisions concerning infrastructure management thus include many facets, viewpoints, e.g. the techniques to employ, current regulations, the impact on user safety, and operating costs, all within a changing social context. This problem raised in assisting decision-makers is therefore essentially multi criteria in nature and dependent upon technical, regulatory, safety and financial aspects. Decisions such as these involve comparing alternatives that have strengths or weaknesses with regard to multiple objectives of interest to the decision maker. Multi-attribute utility theory (MAUT) is a structured methodology designed to handle the tradeoffs among multiple objectives. Utility theory is a systematic approach for quantifying individuals’ preferences. It is used to rescale a numerical value on some measure of interest onto a 0-1 scale with 0 representing the worst preference and 1 the best. This allows the direct comparison of many diverse measures. MAUT can be applied in multidimensional assessments of infrastructures for maintenance or rehabilitation purposes [25], [26]. One of the first applications of MAUT involved a study of alternative locations for a new airport in Mexico City in the early 1970s. The factors that were considered included cost, capacity, access time to the airport, safety, social disruption and noise pollution. As an example, Smith et al. [27] explain that the role of infrastructure management has been continuously changing since the late 1980’s. Indeed, public agencies have started incorporating private sector practises. These new practises include using customer inputs to develop new goals and policies, developing new evaluation procedures for priority programming optimization, and adding feedback loops into infrastructure management systems. One of the new MAUT-based evaluation procedures that has been adopted into infrastructure management is the analytic hierarchy process (AHP). The AHP is a decision making-tool that incorporates both qualitative and quantitative factors. The AHP has increased in use and popularity due to the process reflecting the way people think and make decisions by simplifying complex decisions into a series of one-on-one comparisons [28]. A methodology for prioritising between different maintenance actions in the railway infrastructure is presented in [29]. The consistency of the prioritisation and the feasibility of the applied methodology are investigated. Criteria describing the diverse effects of maintenance are developed and presented to railway systems managers, together with a set of maintenance actions that are specific for each manager. Then, the analytical hierarchy process (AHP) is used to obtain preferences for the criteria and for the different actions. MAUT has also been used in more strategic decision-making modelling such as a dispute resolution selection model prototype for international construction in [30] or a model for the selection of the critical analysis method of construction proposals in a multi actors’ context [31].
Our work deals with MAUT-based multidimensional assessments of infrastructures for maintenance or rehabilitation purposes. The French Public Works Ministry coordinates research projects in the field of engineering infrastructure, particularly within the scope of missions carried out by the Ministry Civil & Urban Engineering Research Network (French acronym: RGC&U). The second point in the 2005 call for research and innovation projects, organized by the French National Research Agency, under the banner “conservation and evaluation of existing facilities”, clearly reveals that decision aid and facilities development constitute key components to the R&D program. Various sponsors of this particular research orientation have set forth an approach built around three distinct and complementary strategies: measurement, evaluation and decision-making. Actors in the area of facility diagnostics and instrumentation (whether researchers or practitioners) are seeking a set of efficient techniques to better identify, qualify and quantify the condition of materials and their flaws. The expert, working as a technical assistant to the project owner, is required to establish instrumentation and assessment strategies while providing a diagnostic and, as such, is seeking access to useful data and procedures in order to generate a reliable diagnostic of the facility. The project owner must ensure that the facility fulfils its functions at optimal cost, within the planned operational framework, and must possess reliable input to enable him to make decisions relative to facility maintenance and repairs, or any necessary structural reinforcement. The various actors are successively involved at three levels: measurement, evaluation, and decision-making. Our works take place in this large-scale project. Our DSS supports the processes of rehabilitation planning and optimisation of the maintenance for the French motorway company ESCOTA.
On January 17, 1956, France first semi-public motorway company was created: ESCOTA (French acronym for the Esterel-Riviera Coast Motorway Company), which today has become a VINCI Concessions1 Group company, whose network coverage is displayed on Fig. 1. The ESCOTA Company has over the past several years built an organization and the accompanying skills to handle all three of these aspects: measurement, evaluation and decision-making. Regarding the measurement and evaluation of the condition of its infrastructure, which comprises 460 kilometres of motorway network comprising 818 bridges, 40 tunnels and 871 operation buildings, the district territorial organisation and ESCOTA Operations Division units strive to ensure that the state of knowledge with respect to facility elements is regularly updated. As for the research efforts associated with the decision-making function, the ESCOTA company decided in 2004 to install a computerized DSS SINERGIE; this system was based on the will for transparency in decision-making, in addition to transferring responsibility to its personnel at each level of intervention and justifying its decision rationale within a multi criteria and multi-actor context [32], [33], [34].
Human–computer interaction and the social implications of computer technology are at the heart of our approach which is supported by a MAUT-based DSS. It first consists in developing models that are predominantly compatible with cognitive modes used by human beings when confronted by a complex situation. Secondly our approach should acknowledge the relations that man or rather an organised group of individuals, establishes with its information system when carrying out actions and making decisions. The research efforts thus insists on knowledge management processes, collective learning and decision-making in the ESCOTA organisation, and then on what sort of information is circulated in such a socio-technical system and how it is approved. A model that explicitly supports the information processing related to the preventive maintenance decisional process of the ESCOTA motorway infrastructure is proposed. This formalization aims at breaking down the barriers existing among actors in ESCOTA through a model of human processes for handling an ill-structured cognitive task: evaluating the ageing motorway infrastructure and identifying the adequate preventive maintenance operations in the Escota organisational hierarchy. Breaking down project evaluation in terms of diagnostic, urgency and priority constitutes a vital information processing phase that facilitates and ultimately prepare a well - substantiated project plan. The multi criteria approach we propose towards infrastructure ageing assessment and decision assistance increases the level of realism and clarity provided to the decision-maker. The aggregation model we propose captures knowledge about the preferences system of ESCOTA’s experts and managers. From a man-machine viewpoint, 1) our decision-making system mainly supports the justification of Escota decision strategy and traces the rationale behind decisions linked to facility management policy utilizing technical, regulatory, safety and financial elements. Indeed, we believe decision elucidation plays a key role in the decision acceptability [35], [33]. That is why our decision-making aid system not only supports the decision-making process itself but also proposes knowledge management opportunities that provide explanations and diagnoses compatible with the cognitive modes of human beings. The ability of a decision support system to justify a decision strategy is an expectation of managers, the DSS must be seen as a recommender system [36], [37], [38], [39]; 2) based upon statistical considerations we also propose robustness analysis techniques of our multi criteria decision-making model that makes ESCOTA engineering decision-making more reliable: the evaluation errors inherent to the infrastructure condition assessments are used to control the decision process and highlight the interactions users can have with the DSS. Indeed, as soon as, an expert doubts the accuracy of an evaluation concerning the condition of the network assets or the manager suspects the accuracy of the diagnosis proposed by the expert on a particular element of the network, further inspections or advanced technical measurements, more quantitative and precise investigations are required to get a more reliable re-evaluation. This refinement of the infrastructure knowledge base is thus controlled through the ESCOTA policy priorities.
This paper is thus a dual-purpose presentation. The ESCOTA Company aims at the formalization and improvement of the decisional process for preventive maintenance in a MC framework. The first purpose is thus to present the main functionalities of the DSS developed to meet the motorway operator requirements. Our mathematical choices have thus been guided by the application constraints. The functionalities of the software are highlighted through a complete and comprehensive example all along the paper. Nevertheless, these functionalities rely on mathematical models that are not of common sense in civil engineering. Part of them refer to well known approaches in the MAUT, the knowledge discovery (KD) or the operational research (OR) communities. They have been adapted or extended to tackle ESCOTA’s specific requirements–we discuss in particular the case of assessments over finite scales in an aggregation procedure; then, they have been integrated into a consistent information processing chain to formalize the whole decisional process for preventive maintenance. The cohabitation and completion of mathematical tools for decision-making in a unique DSS was a major challenge in our approach. The second purpose of the paper is thus to present this consistent and original data processing chain and its associated mathematical models that support the ESCOTA decisional process for preventive maintenance. Some computations or algorithms as the ESCOTA’s strategy elucidation, the MAUT-based ageing infrastructure assessments over finite discrete scales or the risks assessments procedures are original methods; the originality of other computations is related to their integration in a wholly consistent data processing chain (for example the Macbeth method is only a first step in our discrete scales assessment of the infrastructure condition).
The paper is organized as follows. Section 2 proposes a brief description of the multi criteria hierarchical decisional process for maintaining and managing ESCOTA infrastructure facilities. The main concepts and definitions are introduced. The management of infrastructure facilities at ESCOTA presents all the characteristics of an organisational decision-making process. Section 3 is dedicated to the mathematical aspects of our work, in particular the multi criteria aggregation scheme. Some considerations are given about the way continuous cardinal scales are constructed with the ESCOTA operating domain experts. Then, we present how to build a weighted arithmetic mean (WAM) aggregation operator w.r.t. each operating domain, in order to be consistent with the identified scales. The MACBETH method is the support of these first two steps. The problem related to the finite scales, that the experts use when assigning partial scores to a project, is then considered. A method is proposed to ensure a logically sound interface between symbolic assessments and numerical computations in the framework of WAM aggregation. Section 4 is first devoted to a justification procedure of the ESCOTA operation planning policy. Secondly, section 4 proposes a sensitivity analysis to determine the potential causes of overestimation or underestimation in the evaluation process of a project. The risk of making a wrong assessment w.r.t. a project is examined to decide whether undertaking more quantitative and deeper analyses relatively to this project is worth or not. This risk examination provides the criteria that can be considered as the most probable causes for a wrong estimation: this gives indications about further investigations to be carried out. The various information processing phases laid out in SINERGIE are illustrated for road applications, several SINERGIE man/machine interfaces are provided.2 In this paper, we will consider an illustrative example of the A808 highway.3 The functionalities of the software are thus highlighted through a fictitious, but complete and comprehensive example all along the paper. Finally, the advantages and the limits of our approach are discussed.
Section snippets
Characteristics of ESCOTA decisional process for the maintenance of infrastructure
Responsibility for managing engineering facilities consists in ensuring the durability of structures and preserving its quality in order to offer users a high level of service in terms of safety and comfort. This management effort is manifested by very close monitoring and proceeding with regular maintenance, improvements and enhancements introduced on network assets (NA). All NA components are associated with a specific “facility management field”.
The five facility management fields consist of:
The SINERGIE technical solution
Development of the assistance tool called SINERGIE (French acronym for Interactive Evaluation System for the Renovation and Management of ESCOTA’s Infrastructure) aims at providing a significant step forward in facility management capacities at ESCOTA.5
The SINERGIE application comprises implementation of both:
- •
an intranet information system (IS) that handles all information available on the structures and dedicated as ESCOTA’s facility management memory;
Justification
A decision support system must be able to justify a decision strategy, to argue its decisional logic. Decision elucidation plays a key role in the decision acceptability [32], [33], [34], [36], [37].
This function offers explanation elements to the decision-making rationale at each functional level. For each decision threshold in the project evaluation process, this function seeks the dimensions that have exerted a decisive impact on the evaluations. This step entails identifying the criteria
Conclusion
Infrastructure aging is a constant concern for facility managers, who must be in a position to guarantee user safety and effective facility operations over time. Confronted with a massive amount of information to process, it becomes vital to define a durable decision aid strategy that assists managers in both their decision-making and legitimization of investment rationale.
ESCOTA has invested in a multifaceted approach involving: measurement, evaluation and decision-making. The motorway company
Acknowledgements
This Sinergie tool has been developed within the scope of the thesis work performed by Céline Sanchez through the “CIFRE” contract (an industry agreement promoting training through research, with the financial support of the French Research Ministry), which encompasses the LGI2P Laboratory (computer and production engineering), the CMGD Materials Research Center of Alès School of Mines and the ESCOTA motorway company. Special thanks to Brigitte Mahieu and André Nicolas (ESCOTA). The authors
References (65)
- et al.
Soft computing in engineering design – a review
Advanced Engineering in Informatics
(2008) - et al.
An ontological engineering approach for integrating CAD and GIS in support of infrastructure management
Advanced Engineering in Informatics
(2006) - et al.
Dynamic mobile RFID-based supply chain control and management system in construction
Advanced Engineering in Informatics
(2007) - et al.
A nature-inspired decentralized trust model to reduce information unreliability in complex disaster relief operations
Advanced Engineering in Informatics
(2008) - et al.
UtilNets: a water mains rehabilitation decision-support system
Computers Environment and Urban Systems
(2000) Pareto multi-criteria decision making
Advanced Engineering in Informatics
(2008)Elucidative fusion systems: an exposition
Information Fusion
(2000)- et al.
MACBETH – an interactive path towards the construction of cardinal value functions
International Transactions in Operational Research
(1994) - et al.
Quantitative expression and aggregation of performance measurements based on the MACBETH multi-criteria method
International Journal of Production Economics
(2007) - et al.
A review of fuzzy set aggregation connectives
Information Sciences
(1985)
Representation of preferences over a finite scale by a mean operator
Mathematical Social Sciences
Formal order-of-magnitude reasoning in process engineering
Computer and Chemical Engineering
Robust and neutral methods for aggregating preferences into an outranking relation
European Journal of Operational Research
Diagnosis and improvement indexes for a multi-criteria industrial performance synthesized by a Choquet integral aggregation
The International Journal of Management Science, OMEGA
Life cycle impact analysis of energy systems in buildings
Journal of infrastructure Systems
A turnkey client-server decision support architecture for project management
Journal of Decision Systems
Methods for estimating water network rehabilitation needs
Water Supply
Organizing a decision support system for infrastructure maintenance: application to water supply systems
Journal of Decision Systems
Construction integrated management system for contractors
Journal of Buildings and Construction Management
Abstractor: an agglomerative approach to interpreting building monitoring data
ITcon
Towards the next generation of artificial neural networks for civil engineering
Advanced Engineering in Informatics
Cited by (8)
Special issue on RFID and sustainable value chains
2011, Advanced Engineering InformaticsMACBETH
2012, International Journal of Information Technology and Decision MakingOn the mathematical foundations of MACBETH
2016, International Series in Operations Research and Management ScienceA System Supporting the Analysis of Motorway Traffic Accidents
2015, International Journal of Engineering Business ManagementAdaptive Capacity Within a Mega Project: A Case Study on Planning and Decision-Making in the Face of Complexity
2015, European Planning StudiesImproving work bank prioritisation using analytical hierarchical processing
2015, IET Conference Publications