Logistic planning and execution processes in a supply-chain are subject to a high level of complexity because of the number of parties and issues involved, the number of relationships that exist among them, and the dynamic nature of the execution environment. The large volume of data flowing through a sizable computer-based logistic planning and execution management environment that is based on rote data-processing principles tends to overwhelm the human users. As a result many opportunities for improving the efficiency of supply-chain processes and thereby reducing costs are overlooked by the human users, who are forced into a reactive mode. Similar data deluge symptoms are being experienced in other domains such as Internet searches where the number of website hits returned for a single query can easily exceed several million. The data deluge problem could be overcome if the context of the query could be defined by the user and executed by the search engine in a context-based manner. This would require the representation of a virtual model of real world context in the search software. The same need for the representation of context in software exists also in the cyber security domain where data encryption must be supplemented by the profiling of users and the continuous monitoring and automated interpretation of network behavior. This paper discusses the design concepts and implementation principles, and describes the end-state capabilities of a computer-based intelligent logistic planning and execution environment that includes a virtual model of real world supply-chain context and multiple agent groups that are able to interact with each other and the human users. Implemented in a service-oriented architecture (SOA) based infrastructure, the virtual context model provided by a multi-layer ontology and the collaborative agents are able to continuously monitor the state of the supply-chain by interpreting the flow of data in the appropriate context. This allows the agents to rapidly re-plan in case of supply-chain interruptions, discover and act on opportunities for improvements, and identify patterns and trends based on the continuous analysis of historical data. As a result the human users are relieved from lower level data interpretation tasks and provided with actionable information for reactive and proactive planning and execution management functions. The author suggests that order of magnitude improvements in efficiency and reduction in cost are achievable with context-based information-centric software systems.


Software Engineering



URL: https://digitalcommons.calpoly.edu/cadrc/97