Biowaste valorization into lactic acid (LA) by treatment with indigenous microbiota has recently gained considerable attention. LA production from date pulp waste provides an opportunity for resource recovery, reduces environmental issues, and possibly turns biomass into wealth. This study aimed to compare the performance of batch and cyclic fermentation processes in LA production with and without enzymatic pretreatment. The fermentation studies were conducted in the absence of an external inoculum source (relying on indigenous microbiota) and without the addition of nutrients. The highest LA volumetric productivity (3.56 g/liter/day), yield (0.07 g/g-TS), and concentration (21.66 g/L) were attained with enzymatic pretreated date pulp in the cyclic-mode fermentation at the optimized conditions. The productivity rate of LA was enhanced in the cyclic-mode as compared to the batch process. Enzymatic pretreatment increased the digestibility of cellulose that led to higher LA yield. An Artificial Neural Network model was developed to optimize the process parameters and to predict the LA concentration from date pulp waste in both fermentation processes. The main advantage of the ANN approach is the ability to perform quick predictions without resource-consuming experiments. The model predicted optimal conditions well and demonstrated good agreement between experimental and predicted data.

近来，通过用本地微生物菌群处理将生物废物转化为乳酸（LA）的方法受到了广泛关注。从纸浆废物至今的洛杉矶生产为资源回收提供了机会，减少了环境问题，并可能将生物质转化为财富。本研究旨在比较采用和不采用酶预处理的洛杉矶生产中分批发酵和循环发酵过程的性能。发酵研究是在没有外部接种源（依靠本地微生物群）且没有添加营养素的情况下进行的。在优化的条件下，采用酶法预处理的枣粕在循环模式发酵中获得了最高的LA容积生产力（3.56 g /升/天），产量（0.07 g / g-TS）和浓度（21.66 g / L）。与批处理过程相比，LA在循环模式下的生产率得到了提高。酶预处理增加了纤维素的消化率，从而提高了LA产量。开发了一个人工神经网络模型来优化工艺参数并从两个发酵过程中的果肉废料中预测LA的浓度。ANN方法的主要优点是无需消耗资源的实验即可执行快速预测的能力。该模型很好地预测了最佳条件，并证明了实验数据与预测数据之间的良好一致性。开发了一个人工神经网络模型来优化工艺参数并从两个发酵过程中的果肉废料中预测LA的浓度。ANN方法的主要优点是无需消耗资源的实验即可执行快速预测的能力。该模型很好地预测了最佳条件，并证明了实验数据与预测数据之间的良好一致性。开发了一个人工神经网络模型来优化工艺参数并从两个发酵过程中的果肉废料中预测LA的浓度。ANN方法的主要优点是无需消耗资源的实验即可执行快速预测的能力。该模型很好地预测了最佳条件，并证明了实验数据与预测数据之间的良好一致性。