Summary. The natural extension of the simplex method to linear optimization problems with infinitely many constraints applies to their dual problems. Although the feasible sets are convex sets in spaces of generalized finite sequences, they preserve many of the properties of polyhedral convex sets in finite dimensional spaces. These properties are fundamental in obtaining a geometrical interpretation of the pivot operation. The problem of finding a basic set is also analyzed.Zusammenfassung. Die natiifliche Verallgemeinerung des Simplexverfahrens zur Behandlung linearer Optimierungsaufgaben mit unendlich vielen Nebenbedingungen gilt fiir das Duale Problem. Obwohl der zul~issige Bereich zwar konvex im Raume der veraUgemeinerten endlichen Folgen ist, beh/ilt er viele Eigenschaften yore endlichen Fall, die grundlegend um eine geometrische Deutung des Austausch-schritts zu erzeugen sind. Das Problem der Bestimmung yon neuen Basismengen wird ebenfalls behandelt.Any list [7,16] of the existing algorithms for the computational treatment of (P) includes methods with good local convergence [13,14,18], discretizafion techniques [10,17,19,21] and explicit exchange methods [1,7,20,25,26].