Capacitated Lot  Sizing Problems with setup times/costs, inventory, and shortages consideration are the NP-Hard problems requiring high computation resources. Sharma et al. (2017) gave a reformulation of standard Single Level Capacitated Lot Sizing Problem (SLCLSP) that gave the approximate solution in significantly less computation time. They took varying capacity, cost and demand parameters to account uncertainties usually found in the large planning horizon. Though computationally advantageous, their result gave the production plan that exceeded the actual market demand, thus incurring inventory losses. This work uses greedy heuristics to modify the Sharma et al. work such that the level of production matches demand exactly. The heuristic gives better approximate solution (leveled production) without affecting the computational advantage obtained by them.