Towards online planning for open-air engineering processes

Abstract

Recently, most open-air engineering processes, such as open-pit mining, road construction and agriculture farming, are performed with high-tech mobile equipment. The modern technology of the equipment allows optimisation of the operations with respect to the operating conditions. When several mobile equipments are used together in a process, the benefits from their individual optimisation may disappear. Therefore, the execution of an open-air engineering process by cooperating mobile equipment needs to be carefully planned for their effective utilisation. However, dynamics in the open and distributed operating environment of these processes make the planning truly complex. Important planning information, such as the duration of operations, cannot be accurately determined in advance. The progress of the process, equipment performance and operating conditions need to be monitored and this information should be considered for the planning. This paper presents an approach to develop an online planning system with a focus on enhancing process visibility to enable better-informed decision making. To realise these objectives, the PROSA reference architecture and the accompanying delegate multi-agent system are considered. The online planning system based on PROSA is model-driven. It uses executable models to model operations, generating short-term operational forecasts. The forecasting mechanism provides a realistic view on the progress of the process, alerts on potential problems and opportunities and enables to evaluate planning alternatives. This online planning system is implemented as a prototype for open-pit mine planning. However, due to the model-driven nature of the system, this approach is equally applicable to other outdoor engineering processes.

Introduction

Nowadays, operations of open-air engineering processes such as open-pit mining, road construction and agriculture farming are mostly carried out with high-tech mobile equipment. This equipment mainly includes self-propelled work vehicles such as excavators, dump trucks, asphalt layers, road graders, etc. which are designed to carry out specific tasks of the processes. Open-air engineering processes are capital intensive and the operating costs of the work vehicles account for a major proportion of the total process cost. Over the last few years, substantial advancement in the technological development of the work vehicles can be observed. Besides mechanical improvements, an increasing interest is directed in optimising the productivity of the work vehicles through proper planning and execution of their operations (see, e.g. [1], [2], [12]).

Planning for open-air engineering processes involves resource allocation and scheduling decisions, aiming to optimise one or more performance objectives (e.g. minimising completion time, bottleneck utilisation, minimising energy consumption, etc.). In an open-air engineering process, the work vehicles usually perform operations at geographically distributed locations (mine site, storage depot, etc.). The planning for coordinating their operations is necessary for successfully completing the overall process. Because of the open and distributed nature of open-air engineering processes, disturbances and variations are highly prevalent in their operating environments (see, e.g. Table 1). In practice, these processes are typically planned with various project planning methods like Programme Evaluation Review Technique (PERT) and Critical Path Method (CPM) [3], [21]. The plans are generated before the process starts, based on approximate resource performance and predicted operating conditions. Although these plans provide a good starting reference for execution, they are unable to provide the necessary guidance for continued execution of the processes, which are subject to uncertainties and variations. For effective execution, gaining visibility at runtime hence is imperative. Indeed, visibility of possible outcomes of planning choices and visibility of the solution space under the prevailing operating conditions lead to better-informed decision making.

Considering the above requirements, the primary objective of the paper is to present an approach for online planning, which enables to generate plans based on runtime information, monitor the execution and take necessary corrective action when required. As a secondary objective, this paper assesses tests and validates the PROSA reference architecture [4], [5] and the delegate multi-agent system or D-MAS [19] to realise the objectives of online planning.

The PROSA reference architecture provides a basis to design and develop a model-driven system, in which models are executable. The D-MAS provides a coordination pattern to use these executable models in order to virtually execute operations and generate short-term operational forecasts.

A prototype implementation of the online planning approach is carried out for open-pit mine planning. Tests scenarios are used in simulation mode to demonstrate enhanced process visibility, which assists in keeping the plans valid and effective throughout the process for changing conditions over time. The model-driven nature of the system makes it possible to apply this solution in a generic manner to other open-air engineering processes. In particular, these models – driving the system – mirror the operations and their interactions. This effectively encapsulates the specifics of the engineering processes and their environment, which renders the approach generic with respect to open-air application domains. This equally contributes to explaining why PROSA and D-MAS have been able to cope with the novel application area.

The remaining paper is organised as follows. Section 2 elaborates further on the characteristics and the planning requirements of open-air engineering processes. The research is situated within its context and relevant literature is considered in the same section. The online planning method is discussed in Section 3. In Section 4, the prototype implementation of the proposed planning system for open-pit mine planning is described and results are discussed. Finally, Section 5 concludes this paper.