BatchKit is a knowledge integration environment for process engineerin
g. It provides a broad range of problem representation and solving cap
abilities, from trivial calculations and simple inferences to complex
optimization problems that are addressed by the integration of problem
solvers of varying generality such as GAMS (general modelling languag
e and optimization solvers) or SPEEDUP (mass and energy equilibrium an
d dynamic process simulator for chemical processing plants). BatchKit
currently provides prototype conceptual models of plants, processes, p
roduction plans and their various specializations. Representation is b
ased on the integration of object-oriented, logic and constraint progr
amming. The integration of logic programming (LP) supports backtrackin
g search formulations and problem analysis (such as problem-specific c
onsistency checks) using resolution-based inference. The integration o
f constraints as in constraint logic programming (CLP) rounds up the m
odelling capabilities by introducing the equations and inequalities an
d establishes the prerequisites of integrating the special-purpose sol
vers. Human interaction widely relies upon graphic interfaces develope
d and implemented using the graphic building tools of the KEE system b
y IntelliCorp. Knowledge maintenance is facilitated by providing multi
ple external representations for every type of object-graphs, tables a
nd, for more detailed inspection, an assert and query language of the
logic programming. Complete history is maintained to support chronolog
ical backtracking or checkpoint rollback and knowledge state backup. (
C) 1997 Elsevier Science Ltd.