Adaptive geo-information processing service evolution: Reuse and local modification method

Geo-information (GI) service automated composition according to user demands is a crucial task in spatial data infrastructures. State-of-the-art GI service composition approaches face serious limitations in terms of effectiveness and stability as the general GI processing service chain (GIPSC) must be generated from individual user specifications from scratch. This paper presents a novel approach called an adaptive geo-information service evolution (AgiSE) method which overcomes these limitations by adaptively reusing and modifying previously generated GIPSC. In this method, an influence domain minimisation (IDM) criterion is employed to modify the existing GIPSC to fit the new (changed) user demands through minimum revisions. The correction of local modification is ensured by process and integrity constraints. An innovative algorithm called influence domain pursuit is developed to find the optimised solution through a heuristic backward search based on the defined IDM. Experimental analysis shows the significant improvements of using AgiSE in GI services compared with existing traditional methods. The benefits of AgiSE are the improved efficiency of GI service composition and the improved executing stability of GIPSC which were achieved by reducing the service provider load. The AgiSE presented in this paper is crucial in reusing a general unified framework for GI service composition. (ISPRS).