In India and abroad, the commonly used decision modelling in real life rests on the assumption that the decision maker seeks to optimize a well-defined single objective using traditional mathematics programming approach. A farmer may be interested in maximizing his cash income, with certain emphasis on risk minimization. On the other at county level especially in a developing country a planner may aspire for a plan while maximizes food grains production and also to some extent considers employment maximization etc as the goals. Keeping in view the objectives of the study, state-wise secondary data on different variables for the period 1980-81 to 2014-15 were collected from Statistical Abstracts of Punjab, Fertilizer Statistics, Agricultural Statistics at a glance and the reports of the Commission for Agricultural Costs and Prices, published by Ministry of Agriculture By taking its deviations of observed Yt from its estimated value we got the error or the risk coefficients for each year for each crop. These risk coefficients were taken in the matrix formulation in the MOTAD format suggested by Hazell (1971 a and b). To give a meaningful explanation to the level of risk, total mean absolute deviations in gross returns were derived as under: Min A = 1/S Σ│ (chj-gj) xj│ Where A is the minimum average absolute deviation defined as the mean over (h=1………s) years, of the sum of the deviations of gross returns (chj) from the trend in gross returns (gj) multiplied by activity levels x j (j = 1………n). Where A is an unbiased estimator of the population mean absolute income deviation Where A = estimated mean absolute deviation S = no. of years chj = gross returns of the jth activity in hth year gj = sample mean of gross returns of jth activity x j = activity level This was minimized subject to the following constraints: Σaij xj ≤ bi (for all i = 1………….m, j =1……..n) Total activity requirements for the i th constraint, the sum of the unit activity requirements aij for the constraint i times the activity levels ‘xj‘do not exceed the level of the i th constraint bi for all ‘i’ and x j 0 all activity levels are non negative. Where a ij = per unit technical requirement for the jth activity of the ith resource. bi = the ith resource constraint level m = no. of constraints n = no. of activities