Pergamon Expert Systems With Applications, Vol. I I. No. 4, pp. 463473, 1996 Copyright 0 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0957.4174/96 $15.OO+O.O0 PII: SO957-4174(96)00062-O Logistics Information System Auditing Using Expert System Technology I. COMYN-WATTIAU Ecole SupCrieure des Sciences Economiques et Commerciales (ESSEC) and Laboratoire PRiSM, Avenue Bernard Hirsch, BP105.95021 Cergy Cedex, France J. AKOKA ESSEC, Avenue Bernard Hirsch, BP105,95021 Cergy-Pontoise Cedex, France Abstract-This paper brings together two methodological strands of thinking. These are the managerial problem solving methodology of Logistics Information System auditing and the structured development of expert system technology. The investment being made in logistics in organizations is enormous and, although much effort has been devoted to creating structured methods to aid the development of information systems to support these organizations logistics, the area of Logistics Information System auditing remains less developed. The major aim of this paper is to provide a systemic approach of the application of expert system technology to Logistics Information System auditing. Taking a strategic view of Management Information System (MIS) in logistics, this paper describes the application of INFAUDITOR, an audit expert system, to logistics information systems auditing. INFAUDITOR has two fundamental features. First, it covers all domains of information systems, managerial (like logistics) as well as technical aspects. Secondly, it helps to determine, in a given audit situation, the respective importance that should be given to the dcrerent audit domains and tests of control. INFAUDITOR can be viewed as consisting of several expert systems as in blackboard systems. Its fact bases include the characteristics of the enterprise, its logistics information system and the audit objectives. Its rule bases encompass the audit criteria represented as a hierarchical tree. INFAUDITOR is used to assess the ability of a Logistics Information System (LIS) to provide decision makers with relevant, timely information for designing, planning and maintaining an eflcient production system, for securing materials necessary for production, and for facilitating achievement of low operating and maintenance costs. We present and discuss results obtained by using INFAUDITOR in auditing the logistics Management Information System of a large European company. Copyright 0 1996 Elsevier Science Ltd 1. INTRODUCTION THE LAST DECADE has seen an unprecedented rate of development of computer hardware and software, which has created the opportunity for sophisticated data collection, its conversion to meaningful information and the retrieval of that information. During this period, the concept of the Logistics Information System (LIS) has evolved from earlier uses of computers for data process- ing. It is now defined to include everything which deals with the computer-assisted flow and presentation of information, and can be considered to support logistics data-processing functions including Computer Integrated Manufacturing (CIM). Enormous investment is currently being made in LIS and CIM. However, apart from overall cost-benefit analysis, little attention has been paid to assessing the inherent contribution of LIS to an organization (Laudon Akoka Dillard Akoka Felix Fisher, 1984; Hansen Segura, 1985) and for the purpose of developing INFAUDITOR, practicians had to be interviewed to acquire tise. the exper- 3.2. INFAUDITORs Logical Architecture INFAUDITOR can be viewed as consisting of several expert systems as in blackboard systems (Fig. 2). The first knowledge source (KSl) allows the creation of the general audit tree consisting of all the domains and the criteria considered. The second knowledge source (KS2) customizes the general tree on the basis of audit domains. The last knowledge source (KS3) allows an evaluation of the domain considered. The audit tree successively transformed by each KS constitutes the blackboard data. The database, the control data module and KS control are those provided by GURU (MDBS, 1986). In this framework, we do not use different inference engines. However, GURU allows the partition- ing of the knowledge needed to solve the problem into separate and independent KS. For the user interface, we have completed GURU with ACCESS. The audit process is achieved through the following steps: 3.2.1. Step 1. General Audit Tree Generation. The objects on the blackboard are organized as a multi- hierarchical graph. They represent the set of all criteria (auditability, consistency, conformity, efficiency, ability to evolve, feasibility, security, reliability and perform- ance) generally used in the evaluation process and the Blackboard data as a multi-hierarchical graph Control c * Control FIGURE 2. The blackboard architecture (adapted from Engelmore 6 6.1 5 : 6 87 7 7.4 ; 7 7.7 8 7 8 7